Vocabulary knowledge greatly affects writing performance (Stæhr in Lang Learn J 36:139–152, 2008; Johnson in Tesol J 7:700-715 2016), but little is known about the relative contribution of different dimensions of vocabulary knowledge to reading-to-write performance. The current study attempted to investigates the contribution of receptive/orthographic (RecOrth) vocabulary knowledge, productive/orthographic knowledge (ProOrth), productive/phonological (ProPhon) vocabulary knowledge and depth of vocabulary knowledge to reading-to-write scores. For this purpose, 154 Chinese English as foreign language (EFL) learners took a battery of vocabulary knowledge tests and a reading-to-write test. The extent to which vocabulary at different word frequencies predicted reading-to-write was also investigated. The results of regression indicated that ProOrth academic level, vocabulary depth, and RecOrth 2, 000 frequency level explained 40.2% of the reading-to-write score variance. Among the high-performing group, ProOrth academic and vocabulary depth were predictive of the reading-to-write score, while only ProOrth academic vocabulary explained the variance in the reading-to-write score for the low-performing group. The findings reveal the important relationship among dimensions of vocabulary knowledge and reading-to-write and stress the need for systematic vocabulary instruction.