diff --git a/pgcli/pgcompleter.py b/pgcli/pgcompleter.py
index fddd8e7d..bdf71a5b 100644
--- a/pgcli/pgcompleter.py
+++ b/pgcli/pgcompleter.py
@@ -223,7 +223,7 @@ class PGCompleter(Completer):
             pat = re.compile('(%s)' % regex)
 
             def _match(item):
-                r = pat.search(self.unescape_name(item))
+                r = pat.search(self.unescape_name(item.lower()))
                 if r:
                     return -len(r.group()), -r.start()
         else:
diff --git a/tests/test_fuzzy_completion.py b/tests/test_fuzzy_completion.py
index 6b8e6221..6ec8f5f2 100644
--- a/tests/test_fuzzy_completion.py
+++ b/tests/test_fuzzy_completion.py
@@ -74,3 +74,16 @@ def test_should_break_ties_using_lexical_order(completer, collection):
     matches = completer.find_matches(text, collection)
 
     assert matches[1].priority > matches[0].priority
+
+def test_matching_should_be_case_insensitive(completer):
+    """Fuzzy matching should keep matches even if letter casing doesn't match.
+
+    This test checks that variations of the text which have different casing
+    are still matched.
+    """
+
+    text = 'foo'
+    collection = ['Foo', 'FOO', 'fOO']
+    matches = completer.find_matches(text, collection)
+
+    assert len(matches) == 3