This study aims to determine the effectiveness of learning with a scientific approach in the STAD-based cooperative model lesson study of learning outcomes in mathematics. This research is a quasi-experimental study. The study population was junior high school students in Rokan Hilir Regency which were divided into three school levels; upper level, medium level and lower level. The research sample was taken randomly (clustering random sampling). Each sample contained an experimental class and a control class. The research instruments were learning devices and instruments for collecting data. The design of hypothesis testing of student learning outcomes in the form of pretest and posttest data analysis through statistical tests. The data was analyzed using integrated statistical tests. The results showe that student learning outcomes in the experimental class were better than the control class for each level. There are differences in student learning outcomes in the experimental class viewed from the school level, and learning with a scientific approach in cooperative type STAD-based lesson study is effective in terms of student learning outcomes. The conclusion is that the learning with the scientific approach in the STAD-based cooperative model is effective.