The present study aimed to examine whether the first-letter identity/position advantage in the Roman script could be generalized to Chinese-English bilinguals with different scripts and to address the mechanism underlying the first-letter advantage in Chinese-English bilinguals. Four types of nonwords (substituted-first letter nonwords, e.g. lrunk; substituted-internal letter nonwords, e.g. tronk; transposed-first letter nonwords, e.g. rtunk; transposed-internal letter nonwords; e.g. turnk) plus base words (e.g. trunk) were briefly presented to participants, and the EEG signal of participants was recorded while they performed a lexical decision task. ERP results in N400 show that there was a first-letter identity advantage in Chinese-English bilinguals, and the first-letter position advantage increased with English proficiency. These findings suggest that (a) the first-letter advantage also exists in Chinese-English bilinguals, and (b) the mechanism behind it is adaptive modification rather than sequential processing, which supports the modified receptive field (MRF) hypothesis.