The purpose of this study is to describe the ability of students in solving mathematical problems after learning with heuristic strategies. This quasi-experimental study uses a randomize control group pretest-posttest design. The experimental group is a group of students who are taught with heuristic strategies, and the control group is students who are taught with conventional approaches. The population is 280 high school grade X students. Samples n = 74 for the experimental class, and n = 72 for the control class were taken through cluster random sampling. Based on data analysis through analysis of covariant, the results obtained that learning with heuristic strategies are better than conventional approaches to students' ability to solve mathematical problems.