This study is principally conducted as a venture in the domain of vocabulary assessment dominated by multiple-choice, as well as a response to the qualms and suspicion about the feasibility of "Word Association Approach" as an instrument of vocabulary test. Borrowing heavily from the linguistic fields of semantics and lexicology, the author strives for adapting traditional word association approach and eventually comes up with a new version of word association, featuring less freedom in terms of association power and more fitness for Chinese language learners. The real tests and the preceding pilot test aimed to enhance the reliability and validity of the test battery itself yielded abundant information on the examinees as well as on the test itself. Through both quantitative and qualitative data analyses, this study rounds up sufficient evidence to prove not only does word association work as an instrument of vocabulary test, but it works much better than the most popular form, multiple-choice. What’s more, word association stands out more efficiently in probing into the productive vocabulary of English language learners. The high correlation coefficient between word association and the writing part testifies to the closeness of word association to one’s productivity in the English language.