Research on the cognitive neural mechanisms of language control often overlooks the role of rewards. To investigate how reversal rewards affect bilingual language switching during observational learning, we conducted a dual-brain electroencephalography (EEG) study. Participants, classified as direct learners or observers, performed a voluntary language-switching task under dynamic reward conditions. Our results demonstrated that both direct learners and observers exhibited high correct acquisition rates for the switch and non-switch behaviors in both pre- and post-reversal phases. Notably, direct learners and observers showed reduced switch costs in the post-reversal phase, highlighting enhanced language control efficiency. EEG analyses revealed that direct learners exhibited late positive component (LPC) switch costs in both pre- and post-reversal phases, while observers showed LPC switch costs only in the post-reversal phase. These findings support the Adaptive Control Hypothesis by highlighting the adaptability of language control mechanisms in response to dynamic reward environments during direct and observational learning.