This study investigated the effects of different types of captions on English as a Foreign Language Learners' (EFL) vocabulary learning and comprehension. Eighty students in a Chinese university participated. Students were divided into four groups with two classes of freshmen, one class of juniors, and one class of graduate students. Each group watched four video clips with four caption conditions: L1 Chinese, L2 English, dual (L1 and L2), and no captions. The order and caption conditions were counterbalanced. The purpose of the study was to find which caption condition is more effective for EFL learners. Four by four mixed ANOVAs were used to compare the differences among the four conditions and groups. Results indicated that students' performances were statistically significantly different across captions and class levels. In general, students in L1, L2, and dual captions statistically outperformed the no caption condition in vocabulary and comprehension. Results of the effects of L1, L2, and dual captions on vocabulary learning and comprehension were mixed. The pedagogical implications of using authentic TV series and multimedia captions were discussed.