Speech production requires the combined efforts of feedforward and feedback control, but it remains unclear whether the relative weighting of feedforward and feedback control is organized differently between the first language (L1) and the second language (L2). In the present study, a group of Chinese–English bilinguals named pictures in their L1 and L2, while being exposed to multitalker noise. Experiment 1 compared feedforward control between L1 and L2 speech production by examining intensity increases in response to a masking noise (90 dB SPL). Experiment 2 compared feedback control between L1 and L2 speech production by examining intensity increases in response to a weak (30 dB SPL) or strong noise (60 dB SPL). We also examined a potential relationship between L2 fluency and the relative weighting of feedforward and feedback systems. The results indicated that L2 speech production relies less on feedforward control relative to L1, exhibiting attenuated Lombard effects to the masking noise. In contrast, L2 speech production relies more on feedback control than L1, producing larger Lombard effects to the weak and strong noise. The relative weighting of feedforward and feedback control is dynamically changed as second language learning progresses.