Despite the rapid development of the field of Mobile-assisted Language Learning (MALL), research synthesis and systematic meta-analyses on MALL are still lacking. It remains unclear how effective mobile devices are for language learning under different conditions. Review studies on the overall effectiveness of the latest smart mobile devices are still scant. In order to evaluate the learning outcomes of MALL and the impact of moderator variables, we systematically searched journal articles, conference proceedings, and doctoral dissertations published during 2008-2018 and performed a meta-analysis based on a synthesis of 84 effect sizes from 80 experimental and quasi-experimental studies. A medium-to-high effect size of 0.722 was found for the overall effectiveness of using mobile devices for language learning. The findings indicate that the use of mobile devices for language learning is more effective than conventional methods. The effects of nine moderator variables were analyzed. The target language skill, target language and first/second language were found to be significant moderators. Implications for language teaching and research are discussed.