The ability of students in this school is low in understanding the questions given by the teacher. As a result they find it difficult to know the steps on how to solve the problem. Efforts that can be done are implementing Polya's steps.Quasi-experimental research was conducted in this study with the Randomized Control Group Only Design. Students of class VIII at this school are used as population. The experimental class is class VIIIA and the control class is class VIIIB. Final test data in the form of essays obtained were analyzed by t test. The result is the average problem solving ability with the application of Polya's steps, 78.23 and the control class, 72.38, with achievement in the experimental class 81% and in the control class, 48%. After being analyzed, t_count = 2.33 and t_table = 1.68 with a confidence level of 95% so that t_count > t_table (2.33 > 1.68), meaning that the proposed hypothesis is accepted and the ability of students in solving problems using Polya steps is higher rather than without applying these steps.