MOOC allows students to study on-line through Internet at home or some other places, which transcends the barriers of geography and makes the study process easier. As the courses become massive, the problem of information overload plagues students, which makes the course taking difficult to students. Teaching evaluation aims to test the teaching quality of teachers for different courses, which is important to understand the course quality. However, the teaching evaluation of existing MOOC platform is too rough since only five levels are assigned to each course. Besides, there is also a lack of an effective evaluation for teachers. In this paper, we proposed a teaching evaluation model, namely TeachRank, which evaluate teaching quality based on the teaching network. The node in the teaching network represents students, teachers and courses, and the links include the links corresponding to course-taking relationship between students and courses, teaching relationship between teachers and courses. And then, we extend PageRank to teaching network for computing the node quality, in which the scores rated by the students and the link weights are considered. Not only the courses but also the teachers are evaluated by TeachRank. Experiments on real data demonstrate the effectiveness of our proposed TeachRank.