Despite the growing body of research regarding the effectiveness of MALL (mobile-assisted language learning) technologies on foreign language (FL) learners' speaking skill development, a comprehensively quantitative meta-analysis regarding the effect sizes of these studies is still lacking. To solve the problem, this study reported results based on a meta-analysis of 20 effect sizes among 932 participants from 18 experimental and quasi-experimental studies. The results showed that the overall effect size was significantly large, suggesting the use of MALL for FL learners' speaking skill development is more effective than traditional methods. Furthermore, learner-related, instruction-related and methodology-related moderator analysis results indicated that instructional approaches and intervention durations were significant moderators, while proficiency levels, educational levels, language types, intervention settings, software types, measured outcome types and duration intensity did not find a significant moderating effect. The results of the study provide some pedagogical implications into the use of MALL technologies for FL learners' speaking skill development.