This study examines how visual information in nonnative speech affects native listener judgments of second language (L2) speech production. Native Canadian English listeners perceived three English phonemic contrasts (/b-v, θ-s, l-ɹ/) produced by native Japanese speakers as well as native Canadian English speakers as controls. Among the stimuli, /v, θ, l, ɹ/ are not existent in the Japanese consonant inventory. These stimuli were presented under audio-visual (AV), audio-only (AO), and visual-only (VO) conditions. The results showed that while overall perceptual judgments of the nonnative phonemes (/v, θ, l, ɹ/) were significantly less intelligible than the native phonemes (/b,s/), the English listeners perceived the Japanese productions of the phonemes /v, θ, b,s/ as significantly more intelligible when presented in the AV condition compared to the AO condition. However, the Japanese production of /ɹ/ was perceived as less intelligible in the AV compared to the AO condition. Further analysis revealed that a significant number of Japanese productions of /ɹ/ lacked lip-rounding, indicating that nonnative speakers’ incorrect articulatory configurations may decrease intelligibility. These results suggest that visual cues in L2 speech productions may be either facilitative or inhibitory in native perception of L2 accented-speech. [Research supported by SFU and SSHRC.]