diff --git a/moto/ec2/models.py b/moto/ec2/models.py
index b2e6b050e..6f9ada1f3 100644
--- a/moto/ec2/models.py
+++ b/moto/ec2/models.py
@@ -70,7 +70,7 @@ from .utils import (
     random_volume_id,
     random_vpc_id,
     random_vpc_peering_connection_id,
-    is_filter_matching)
+    generic_filter)
 
 
 class InstanceState(object):
@@ -96,6 +96,9 @@ class TaggedEC2Instance(object):
         if filter_name == 'tag-key':
             return [tag['key'] for tag in tags]
 
+        if filter_name == 'tag-value':
+            return [tag['value'] for tag in tags]
+
 
 class NetworkInterface(object):
     def __init__(self, subnet, private_ip_address, device_index=0, public_ip_auto_assign=True, group_ids=None):
@@ -1156,7 +1159,7 @@ class VPC(TaggedEC2Instance):
 
         filter_value = super(VPC, self).get_filter_value(filter_name)
 
-        if not filter_value:
+        if filter_value is None:
             msg = "The filter '{0}' for DescribeVPCs has not been" \
                   " implemented in Moto yet. Feel free to open an issue at" \
                   " https://github.com/spulec/moto/issues".format(filter_name)
@@ -1195,11 +1198,7 @@ class VPCBackend(object):
         else:
             vpcs = self.vpcs.values()
 
-        if filters:
-            for (_filter, _filter_value) in filters.items():
-                vpcs = [ vpc for vpc in vpcs if is_filter_matching(vpc, _filter, _filter_value) ]
-
-        return vpcs
+        return generic_filter(filters, vpcs)
 
     def delete_vpc(self, vpc_id):
         # Delete route table if only main route table remains.
diff --git a/moto/ec2/utils.py b/moto/ec2/utils.py
index a6935aa2c..8a0c702d3 100644
--- a/moto/ec2/utils.py
+++ b/moto/ec2/utils.py
@@ -286,11 +286,19 @@ def is_filter_matching(obj, filter, filter_value):
 
     try:
         value = set(value)
-        return value.issubset(filter_value) or value.issuperset(filter_value)
+        return (value and value.issubset(filter_value)) or value.issuperset(filter_value)
     except TypeError:
         return value in filter_value
 
 
+def generic_filter(filters, objects):
+    if filters:
+        for (_filter, _filter_value) in filters.items():
+            objects = [obj for obj in objects if is_filter_matching(obj, _filter, _filter_value)]
+
+    return objects
+
+
 # not really random ( http://xkcd.com/221/ )
 def random_key_pair():
     return {
diff --git a/tests/test_ec2/test_vpcs.py b/tests/test_ec2/test_vpcs.py
index 47dc855c4..9f3bc8351 100644
--- a/tests/test_ec2/test_vpcs.py
+++ b/tests/test_ec2/test_vpcs.py
@@ -150,6 +150,7 @@ def test_vpc_get_by_tag_key_superset():
     vpc1.id.should.be.within(vpc_ids)
     vpc2.id.should.be.within(vpc_ids)
 
+
 @mock_ec2
 def test_vpc_get_by_tag_key_subset():
     conn = boto.connect_vpc()
@@ -167,4 +168,43 @@ def test_vpc_get_by_tag_key_subset():
     vpcs.should.have.length_of(2)
     vpc_ids = tuple(map(lambda v: v.id, vpcs))
     vpc1.id.should.be.within(vpc_ids)
+    vpc2.id.should.be.within(vpc_ids)
+
+
+@mock_ec2
+def test_vpc_get_by_tag_value_superset():
+    conn = boto.connect_vpc()
+    vpc1 = conn.create_vpc("10.0.0.0/16")
+    vpc2 = conn.create_vpc("10.0.0.0/16")
+    vpc3 = conn.create_vpc("10.0.0.0/24")
+
+    vpc1.add_tag('Name', 'TestVPC')
+    vpc1.add_tag('Key', 'TestVPC2')
+    vpc2.add_tag('Name', 'TestVPC')
+    vpc2.add_tag('Key', 'TestVPC2')
+    vpc3.add_tag('Key', 'TestVPC2')
+
+    vpcs = conn.get_all_vpcs(filters={'tag-value': 'TestVPC'})
+    vpcs.should.have.length_of(2)
+    vpc_ids = tuple(map(lambda v: v.id, vpcs))
+    vpc1.id.should.be.within(vpc_ids)
+    vpc2.id.should.be.within(vpc_ids)
+
+
+@mock_ec2
+def test_vpc_get_by_tag_value_subset():
+    conn = boto.connect_vpc()
+    vpc1 = conn.create_vpc("10.0.0.0/16")
+    vpc2 = conn.create_vpc("10.0.0.0/16")
+    vpc3 = conn.create_vpc("10.0.0.0/24")
+
+    vpc1.add_tag('Name', 'TestVPC')
+    vpc1.add_tag('Key', 'TestVPC2')
+    vpc2.add_tag('Name', 'TestVPC')
+    vpc2.add_tag('Key', 'TestVPC2')
+
+    vpcs = conn.get_all_vpcs(filters={'tag-value': ['TestVPC', 'TestVPC2']})
+    vpcs.should.have.length_of(2)
+    vpc_ids = tuple(map(lambda v: v.id, vpcs))
+    vpc1.id.should.be.within(vpc_ids)
     vpc2.id.should.be.within(vpc_ids)
\ No newline at end of file