diff --git a/word_way/scrapping/word.py b/word_way/scrapping/word.py
index 7c2bf81..87ea545 100644
--- a/word_way/scrapping/word.py
+++ b/word_way/scrapping/word.py
@@ -16,7 +16,7 @@
 from word_way.models import Pronunciation, Sentence, Word, WordSentenceAssoc
 from word_way.utils import convert_word_part
 
-__all__ = 'save_word', 'save_word_task',
+__all__ = 'save_word', 'save_word_task', 'save_words_task',
 
 
 @celery.task
@@ -24,6 +24,20 @@ def save_word_task(target_word: str):
     save_word(target_word, session)
 
 
+@celery.task
+def save_words_task():
+    """단어가 없는 발음을 가져와서 단어를 저장하는 테스크"""
+    q = session.query(Word).filter(Word.pronunciation_id == Pronunciation.id)
+    pronunciations = session.query(Pronunciation).filter(
+        ~q.exists()
+    ).all()
+    for pronunciation in pronunciations:
+        pronunciation_id = save_word(pronunciation.pronunciation, session)
+        if not pronunciation_id:
+            session.delete(pronunciation)
+            session.commit()
+
+
 def save_word(
     target_word: str, session: Session,
 ) -> typing.Optional[uuid.UUID]: