While researchers in the second language (L2) field often consider that L2 anxiety determines subsequent L2 achievement, an emerging line of research suggests that language skills better predict L2 anxiety. This viewpoint has yet to be experimentally evaluated, and thus it motivated the present study. Two groups of university language students enrolled in the Department of English as an L2 were followed over one semester at three time points while taking a general university course unrelated to language learning. The experimental group received extra instruction (course unrelated) designed to expand their L2 vocabulary knowledge, including strategies for learning and using new vocabulary in real-life contexts over the whole semester. A typical teaching method was delivered to the control group. The conditional dual-domain latent growth curve modeling (LGCM) with grouping variable as a covariate was used to study the possible effect of the intervention on the trajectory of language anxiety and vocabulary knowledge. Results indicated that teaching vocabulary knowledge to the experimental group led to an increase in their L2 vocabulary achievement. In turn, the increase in L2 vocabulary significantly reduced students' L2 anxiety. Notably, this reduction in anxiety was attributed solely to the improvement in vocabulary achievement, as teachers did not employ any anxiety-controlling strategies. In contrast, the control group showed no significant changes in L2 vocabulary knowledge or L2 anxiety, which aligns with typical observations in standard learning settings. Findings from the present experimental study support the idea that increasing students' L2 achievement (e.g. vocabulary) through providing practical resources, strategies, and opportunities for using the language (even with minor errors) can reduce their L2 anxiety. According to our findings, using language-enhancing strategies, rather than anxiety-reducing ones, can more effectively help reduce L2 anxiety.