Although there were studies that presented the applications of metacognitive skill training, the research on web-based metacognitive skills training are few. The purpose of this study is to design a web-based learning environment and further examine the effect of the web-based training. A pretest-posttest quasi-experimental design was used in this study. Fifty-three college students were assigned into experimental and control groups. After four-week training period, the results of paired-samples t-test showed that experimental group's posttest scores were significantly higher than the pretest scores in self-plan, self-monitor, and total score, while there was no significance in the control group. In addition, students in experimental group made significantly greater gains compared to control group in self-plan. Discussion and suggestions are also provided. (Contains 2 tables and 3 figures.)