As a learner-centered approach, technology-facilitated personalized learning has received increasing attention in the field of education. However, it is not clear whether technology-facilitated personalized learning is effective for both learning achievements and learning perceptions. This study aims to close the research gaps and examine the effect sizes of technology-facilitated personalized learning on learning achievements and learning perceptions from 2001 to 2020. In total, 34 empirical studies were synthesized in the present meta-analysis. The findings revealed that technology-facilitated personalized learning had a medium effect size on learning achievements and a small effect size on learning perceptions. Furthermore, the impacts of 10 moderator variables, including the sample levels, sample sizes, learning domains, research design, research settings, intervention duration, learning methods, software and hardware that support personalized learning, and user-oriented personalized parameters, were analyzed in depth. It was found that learning methods and personalized learning software significantly moderated the effect sizes. The novelty and main contribution of this study is threefold. First, this study revealed that technology-facilitated personalized learning was more effective than traditional learning through the comprehensive meta-analysis. Second, this study found that learning methods and personalized learning software could trigger and leverage the effects of technology-facilitated personalized learning. Third, this study proposed a design framework of technology-facilitated personalized learning to facilitate the implementation of personalized learning effectively.