Web applications have gained widespread use in various contexts, and their increasing popularity has led to growing demands for automated testing. Testing Web applications requires operating Web browsers based on test scenarios, which can be costly work for test engineers. Moreover, there can be various negative test cases corresponding to operational errors caused by human errors, and thorough testing becomes more challenging. This paper proposes a novel method for generating negative test cases from normal test cases by utilizing the mutation technique. Furthermore, the paper proposes categorizing negative test cases by their final screen (i.e., the last Web page) similarity to help check their behaviors efficiently. The paper then examines the usefulness of the proposed method through an experiment.