This research aims to analyze the results of the application of Blended Learning with reference to the Project-Based Learning (PjBL) model. Blended Learning and PjBL can be packaged as a learning model, namely Project-Based Blended Learning (PjB2L). The design of learning activities with the help of e-campus media is used to measure its impact on student learning outcomes in Statistics courses. The main focus of the research describes two main things, the description of the application of statistics lectures with the PjB2L model and the results of statistical tests on differences in learning outcomes obtained. Quasi-experimental research was designed to compare student learning outcomes in experimental classes (with the PjB2L model) and control classes (conventional/blended direct learning). 43 students were involved in this study which was divided into two classes, namely the control class which included 20 students, and the experimental class which involved 23 students. In addition, on data analysis technique, analysis of covariance (ANCOVA), was used to compare the results of the control group and the experimental group, with the covariance variable being the pre-test results (initial knowledge). The results showed significant differences. The difference shown in this study leads to the support of the experimental group's better learning achievement.