As gamification is introduced into education, researchers believe it has the potential to improve online learning. However, the effects of gamified online learning (GOL) are mixed. Is it more effective than online learning? Given that no meta-analysis analyzes the effectiveness of GOL. This meta-analysis explores the effects of GOL on students' learning based on 27 experimental studies. The results show that GOL has a medium positive effect on students' learning (SMD=0.533, 95% CI [0.405, 0.662], p<0.001), especially on academic achievement (SMD=0.658, p<0.001). Furthermore, moderator analyses show that the learning outcome, educational level, group activity, competition or cooperation, game element, research design, and region moderate the effects of GOL. GOL has better effects under the below conditions: (1) class sizes are 31-50; (2) treatment duration is "three" months or less; (3) among university and in-service students; (4) cooperation or mixed (cooperation + competition) are better; (5) group + individual learning is better; (6) except for progress bars, other "nine" game elements can promote students' learning; (7) four or fewer types of game elements. Finally, discussion, implications, and conclusion are put forward.