The first step in writing a math test paper is to choose the right topic. Now I just have the ability to choose good questions, but I don't have the ability to make good questions myself. So according to the two-way detailed list I made, I first selected a number of questions for each knowledge point, which may be easy, medium and difficult, and then selected a test question according to the difficulty of my design. Every knowledge point repeats this process, and then a set of questions is initially determined. Then I made it myself and felt it. Are there some knowledge points that appear too frequently in the questions where some knowledge points meet? If there is, choose a more suitable one from those selected alternative questions, and then see if there is a gradient, and try to make the whole test paper distribute from easy to difficult, but sometimes I like to write questions. I think it is necessary to set a difficult problem in a certain position to examine students' psychological quality and coping strategies in the examination room. In the choice of topics, we must be conducive to the examination of what we have learned. If you often take many questions in exams, you will find that students have been puzzled, and that kind of question will become my must. Of course, if you find that it has been bad, you should also carry out special training in a targeted manner. According to the mistakes made by students in the past, and my understanding of students, I also pay attention to choosing simple and easy-to-make questions when choosing topics, so that students can better understand their mistakes and improve the accuracy. I also like to choose some low-level mistakes because students don't carefully examine the questions, so that students can realize that it is easy to lose points if they don't look at the questions carefully. I also like to write questions that are conducive to students' individuality and innovation. Let the questions not be too rigid and leave room for students to display. I advocate novelty, ingenuity and vividness, but I don't advocate making problems mysterious, strange and biased. Of course, these questions I like generally don't appear too many at the same time on a test paper, otherwise students may have unimaginable consequences when answering them.