diff --git a/examples/flask_sqlalchemy/schema.py b/examples/flask_sqlalchemy/schema.py
index 9ed09464..b509d1f7 100644
--- a/examples/flask_sqlalchemy/schema.py
+++ b/examples/flask_sqlalchemy/schema.py
@@ -3,10 +3,26 @@
 from models import Role as RoleModel
 
 import graphene
+from graphql_relay.node.node import from_global_id
 from graphene import relay
 from graphene_sqlalchemy import SQLAlchemyConnectionField, SQLAlchemyObjectType
 
 
+# Just a util function 
+
+def input_to_dictionary(input):
+    """Method to convert Graphene inputs into dictionary"""
+    dictionary = {}
+    for key in input:
+        # Convert GraphQL global id to database id
+        if key[-2:] == 'id':
+            input[key] = from_global_id(input[key])[1]
+        dictionary[key] = input[key]
+    return dictionary
+
+
+
+# Queries Part :
 class Department(SQLAlchemyObjectType):
     class Meta:
         model = DepartmentModel
@@ -24,7 +40,6 @@ class Meta:
         model = RoleModel
         interfaces = (relay.Node, )
 
-
 class Query(graphene.ObjectType):
     node = relay.Node.Field()
     # Allow only single column sorting
@@ -35,5 +50,25 @@ class Query(graphene.ObjectType):
     # Disable sorting over this field
     all_departments = SQLAlchemyConnectionField(Department, sort=None)
 
+# Mutation Part :
+# only one exemple : employee.
+class CreateEmployeeInput(SQLAlchemyInputObjectType):
+    class Meta:
+        # You have to exclude those fields to avoid conflict
+        exclude_fields = ('id','uuid')
+        model = EmployeeModel
+        
+class CreateEmployee(graphene.Mutation):
+    class Arguments:
+        input = CreateEmployeeInput(required=True)
+        
+    def mutate(self,info,input):
+        data = utils.input_to_dictionary(input)
+        employee = EmployeeModel(**data)
+
+        return CreateEmployee(employee=employee)
 
-schema = graphene.Schema(query=Query, types=[Department, Employee, Role])
+class Mutation(graphene.ObjectType):
+    createEmplyee = CreateEmployee.Field()  
+    
+schema = graphene.Schema(query=Query,mutation=Mutation types=[Department, Employee, Role])
diff --git a/graphene_sqlalchemy/__init__.py b/graphene_sqlalchemy/__init__.py
index d328304a..e9aec9f9 100644
--- a/graphene_sqlalchemy/__init__.py
+++ b/graphene_sqlalchemy/__init__.py
@@ -1,4 +1,4 @@
-from .types import SQLAlchemyObjectType
+from .types import SQLAlchemyObjectType, SQLAlchemyInputObjectType
 from .fields import SQLAlchemyConnectionField
 from .utils import get_query, get_session
 
@@ -7,6 +7,7 @@
 __all__ = [
     "__version__",
     "SQLAlchemyObjectType",
+    "SQLAlchemyInputObjectType",
     "SQLAlchemyConnectionField",
     "get_query",
     "get_session",
diff --git a/graphene_sqlalchemy/types.py b/graphene_sqlalchemy/types.py
index c20e8cfc..411104e0 100644
--- a/graphene_sqlalchemy/types.py
+++ b/graphene_sqlalchemy/types.py
@@ -210,3 +210,69 @@ def enum_for_field(cls, field_name):
     sort_enum = classmethod(sort_enum_for_object_type)
 
     sort_argument = classmethod(sort_argument_for_object_type)
+    
+def construct_fields_for_input(
+    obj_type, model, registry, connection_field_factory,only_fields, exclude_fields
+):
+    inspected_model = sqlalchemyinspect(model)
+
+    fields = OrderedDict()
+
+    for name, column in inspected_model.columns.items():
+        is_not_in_only = only_fields and name not in only_fields
+        # is_already_created = name in options.fields
+        is_excluded = name in exclude_fields  # or is_already_created
+        if is_not_in_only or is_excluded:
+            # We skip this field if we specify only_fields and is not
+            # in there. Or when we exclude this field in exclude_fields
+            continue
+        converted_column = convert_sqlalchemy_column(column, registry)
+        fields[name] = converted_column
+
+    for name, composite in inspected_model.composites.items():
+        is_not_in_only = only_fields and name not in only_fields
+        # is_already_created = name in options.fields
+        is_excluded = name in exclude_fields  # or is_already_created
+        if is_not_in_only or is_excluded:
+            # We skip this field if we specify only_fields and is not
+            # in there. Or when we exclude this field in exclude_fields
+            continue
+        converted_composite = convert_sqlalchemy_composite(composite, registry)
+        fields[name] = converted_composite
+
+    for hybrid_item in inspected_model.all_orm_descriptors:
+
+        if type(hybrid_item) == hybrid_property:
+            name = hybrid_item.__name__
+            is_not_in_only = only_fields and name not in only_fields
+            # is_already_created = name in options.fields
+            is_excluded = name in exclude_fields  # or is_already_created
+            if is_not_in_only or is_excluded:
+                # We skip this field if we specify only_fields and is not
+                # in there. Or when we exclude this field in exclude_fields
+                continue
+            converted_hybrid_property = convert_sqlalchemy_hybrid_method(hybrid_item)
+            
+            fields[name] = converted_hybrid_property
+
+    
+
+    return fields
+
+
+
+class SQLAlchemyInputObjectType(graphene.InputObjectType):
+    @classmethod
+    def __init_subclass_with_meta__(cls, model=None, registry=None, skip_registry=False,only_fields=(), exclude_fields=(),
+                                    connection=None,_meta=None,connection_field_factory=default_connection_field_factory,
+                                    use_connection=None, interfaces=(), id=None, **options):
+
+        if not registry:
+            registry = get_global_registry()
+            
+
+        sqla_fields = yank_fields_from_attrs(
+            construct_fields_for_input(model=model, registry=registry,only_fields=only_fields, exclude_fields=exclude_fields,
+                            connection_field_factory=connection_field_factory,obj_type=cls),
+            _as=graphene.Field,
+        )