This meta-analysis study aimed to scrutinize the effect of the problem-based learning (PBL) approach on students' academic achievement in science lessons compared to the traditional instructional approach. As research data, the researcher accessed 55 experimental studies as a result of steps of a systematic review with the keywords "problem-based learning," "science," and "academic achievement or success" from the Education Resources Information Center, Scopus, Web of Science, Proquest, Taylor and Francis, Center of Higher Education National Thesis Center, and Wiley Online databases. The results showed that the PBL method used in science education had a high effect size on academic achievement compared to the traditional instructional method. The results also explained significant differences between the effect sizes of publication types (conference papers, articles, master's thesis, and doctoral thesis). On the other hand, they demonstrated that academic achievement was similar in the nationality of the studies, educational levels, instructional variations with PBL, and relevant science areas. It is expected that integrating innovative instructional approaches into PBL could guide researchers who establish a research problem about how it can affect science academic achievement.