Video lectures are becoming an important and accessible way for adults to learn a new second language (L2). Unlike face-to-face classroom teaching, teachers are more challenged by the asynchronous mode of video teaching. Finding more effective teaching methods is a great concern for teachers. The present study examined adults' second language (L2) vocabulary learning in a video-based social learning environment. Participants included 116 Chinese undergraduates. They learned German words through one of three pedagogical video lecture conditions: social-context (SC) condition, picture-phonics (PP) condition, and non-picture phonics (NPP) condition. We measured learners' learning performance, cognitive load, intrinsic motivation, and learning satisfaction. The results indicated that adult learners in the SC condition performed significantly better in word meaning retention, had significantly higher intrinsic motivation, and perceived significantly lower cognitive load as compared to the PP and NPP conditions. The implications are that teachers could use social learning pedagogy to create a better video-based learning environment for L2 adult learners for better learning performance.