Although digital game-based vocabulary learning (DGBVL) has received increasing attention in the past two decades, the impacts of DGBVL on the depth of word knowledge are still not well understood, especially in regard to productive vocabulary learning and DGBVL's long-term efficacy. This study leverages a quasiexperimental research design to investigate DGBVL's long-term effects on receptive vocabulary (RV) and productive vocabulary (PV). Forty-eight Chinese English-as-a-foreign-language (EFL) university students, assigned to the experimental and control groups, were instructed by a DGBVL approach and PowerPoint (PPT) lecturing, respectively, over the course of 18 weeks. Specifically, a mixed 2×2 repeated measures experimental design was conducted by adopting instruction type (DGBVL and PPT lecturing) and testing time (pretest and posttest) as the independent variables, with RV and PV proficiency as the respective dependent variables. The results suggest that instruction type and teaching time have significant effects on participants' RV and PV learning achievements. However, teaching time's effect size outweighs instruction type. The findings are highly encouraging for the use of DGBVL in the EFL classroom, as it may serve as an effective and long-lasting pedagogical tool within this context.