diff --git a/docs/api-guide/serializers.md b/docs/api-guide/serializers.md
index 95703033a..5aae96a9f 100644
--- a/docs/api-guide/serializers.md
+++ b/docs/api-guide/serializers.md
@@ -570,6 +570,29 @@ This option is a dictionary, mapping field names to a dictionary of keyword argu
             user.save()
             return user
 
+
+## Return errors on undefined fields
+
+By default, if the user consuming the your API supplies fields in a payload that you have not defined (eg. fields in a POST payload), Django Rest Framework will ignore them.
+
+It is possible to have DRF return validation errors for undefined fields supplied by the user. This is done by setting `error_on_undefined=True` in your Meta class. For example:
+
+    class NameSerializer(serializers.Serializer):
+        first_name = serializers.CharField()
+        last_name = serializers.Charfield()
+
+        class Meta:
+            error_on_undefined = True
+
+    serializer = NameSerializer(data={'first_name': 'George'
+                                      'middle_name': 'Walker',
+                                      'last_name': 'Bush'})
+    serializer.is_valid()  # False
+    print(serializer.errors)  # {'middle_name': 'Not a supported field.'}
+
+This can be helpful to users of your API as they may not realise that they are supplying an unsupported field. For example they may have a typo in their integration code.
+
+
 ## Relational fields
 
 When serializing model instances, there are a number of different ways you might choose to represent relationships.  The default representation for `ModelSerializer` is to use the primary keys of the related instances.
diff --git a/rest_framework/serializers.py b/rest_framework/serializers.py
index 5aef1df69..7f3f94486 100644
--- a/rest_framework/serializers.py
+++ b/rest_framework/serializers.py
@@ -323,7 +323,8 @@ def get_validation_error_detail(exc):
 @six.add_metaclass(SerializerMetaclass)
 class Serializer(BaseSerializer):
     default_error_messages = {
-        'invalid': _('Invalid data. Expected a dictionary, but got {datatype}.')
+        'invalid': _('Invalid data. Expected a dictionary, but got {datatype}.'),
+        'undefined': _('Not a supported field.')
     }
 
     @property
@@ -446,6 +447,13 @@ class Serializer(BaseSerializer):
             else:
                 set_value(ret, field.source_attrs, validated_value)
 
+        if getattr(self, 'Meta', None) and \
+                getattr(self.Meta, 'error_on_undefined', False):
+            undefined_fields = set(data) - set(self.fields)
+            if undefined_fields:
+                errors.update({field: self.error_messages['undefined']
+                               for field in undefined_fields})
+
         if errors:
             raise ValidationError(errors)
 
diff --git a/tests/test_serializer.py b/tests/test_serializer.py
index 741c6ab17..b8aa61f78 100644
--- a/tests/test_serializer.py
+++ b/tests/test_serializer.py
@@ -61,6 +61,18 @@ class TestSerializer:
         with pytest.raises(AssertionError):
             serializer.save()
 
+    def test_error_on_undefined(self):
+        class ExampleSerializer(serializers.Serializer):
+            char = serializers.CharField()
+
+            class Meta():
+                error_on_undefined = True
+
+        serializer = ExampleSerializer(data={'char': 'abc', 'foo': 'bar'})
+        assert not serializer.is_valid()
+        assert serializer.validated_data == {}
+        assert serializer.errors == {'foo': 'Not a supported field.'}
+
 
 class TestValidateMethod:
     def test_non_field_error_validate_method(self):