diff --git a/language/google/cloud/language/syntax.py b/language/google/cloud/language/syntax.py
index 0c7260839e74..e2b4ab286f24 100644
--- a/language/google/cloud/language/syntax.py
+++ b/language/google/cloud/language/syntax.py
@@ -18,6 +18,8 @@
 breaks a document down into tokens and sentences.
 """
 
+from google.cloud.language.sentiment import Sentiment
+
 
 class PartOfSpeech(object):
     """Part of speech of a :class:`Token`."""
@@ -183,11 +185,19 @@ class Sentence(object):
     :param begin: The beginning offset of the sentence in the original
                   document according to the encoding type specified
                   in the API request.
+
+    :type sentiment: :class:`~google.cloud.language.sentiment.Sentiment`
+    :param sentiment:
+        (Optional) For calls to
+        :meth:`~google.cloud.language.document.Document.annotate_text` where
+        ``include_sentiment`` is set to true, this field will contain the
+        sentiment for the sentence.
     """
 
-    def __init__(self, content, begin):
+    def __init__(self, content, begin, sentiment=None):
         self.content = content
         self.begin = begin
+        self.sentiment = sentiment
 
     @classmethod
     def from_api_repr(cls, payload):
@@ -200,4 +210,11 @@ def from_api_repr(cls, payload):
         :returns: The sentence parsed from the API representation.
         """
         text_span = payload['text']
-        return cls(text_span['content'], text_span['beginOffset'])
+
+        try:
+            sentiment = Sentiment.from_api_repr(payload['sentiment'])
+        except KeyError:
+            sentiment = None
+
+        return cls(text_span['content'], text_span['beginOffset'],
+                   sentiment=sentiment)
diff --git a/language/unit_tests/test_syntax.py b/language/unit_tests/test_syntax.py
index da6f09e54de4..b09bc2ce7004 100644
--- a/language/unit_tests/test_syntax.py
+++ b/language/unit_tests/test_syntax.py
@@ -125,3 +125,28 @@ def test_from_api_repr(self):
         sentence = klass.from_api_repr(payload)
         self.assertEqual(sentence.content, content)
         self.assertEqual(sentence.begin, begin)
+        self.assertEqual(sentence.sentiment, None)
+
+    def test_from_api_repr_with_sentiment(self):
+        from google.cloud.language.sentiment import Sentiment
+        klass = self._get_target_class()
+        content = 'All the pretty horses.'
+        begin = -1
+        score = 0.5
+        magnitude = 0.5
+        payload = {
+            'text': {
+                'content': content,
+                'beginOffset': begin,
+            },
+            'sentiment': {
+                'score': score,
+                'magnitude': magnitude,
+            }
+        }
+        sentence = klass.from_api_repr(payload)
+        self.assertEqual(sentence.content, content)
+        self.assertEqual(sentence.begin, begin)
+        self.assertIsInstance(sentence.sentiment, Sentiment)
+        self.assertEqual(sentence.sentiment.score, score)
+        self.assertEqual(sentence.sentiment.magnitude, magnitude)