Abstract: This paper aims to explore the Moodle-based man-machine interaction translation teaching model. The subjects were the college English majors of two natural classes taught by the researcher of the present study. The research tools were the Moodle-based TTP-CLEO designed and developed by the researcher himself, the pretests and posttests, and the SPSS. Frequencies of nine specific translating techniques were classified based on the subjects'translations for comparative analyses in use of Descriptive Statistics and One-Way ANOVA. The results show that there are significant differences between the overall frequencies of the translating techniques employed by the participants in the pretests and posttests, as well as big differences between the frequencies of the participants' individual techniques used. Furthermore, the frequencies of the experimental class are apparently bigger than those of the controlled class, demonstrating that the former reaches the level of preliminary use, while the latter the level betweennonuseandpreliminary use. This Moodle-based man-machine interaction translation teaching model rarely seen in China is proved to help break various limitations of traditional translation teaching, doubtlessly as an active attempt and exploration.