Timetabling problems is one of very significant problems in many fields of applications. As mentioned in [3], this problem is NP-complete with numerous factors and constraints. In general, it is treated as a multi-objective optimization problem. Currently, the work of scheduling is very difficult in universities, especially in credits training. Almost it is impossible to control all cases of the problem by human. Therefore, we cannot manually give an effective solution for this problem. There are quite many methods to resolve this problem in literature, they are mostly searching methods based on genetic algorithms and their results are proved effective in practice. In this paper, we propose a method based on genetic algorithms for university course timetabling problems with some modifications and apply it to real-world datasets in Hanoi Open University.