The objectives of this study are (1) to test the mathematical problem solving ability with RME learning model assisted by Jarimatika better than the mathematical problem solving ability with conventional learning, (2) to test the mathematical problem solving ability with RME model assisted by Jarimatika better than the mathematical problem solving ability with conventional learning, (3) to test the improvement of mathematical problem solving ability with RME model assisted by Jarimatika better than the improvement of mathematical problem solving ability with conventional learning. This research is a pure experimental research. The design was pre test post test control group design. The population was fifth grade students at Jatilawang 1 State Elementary School and Kemantran 1 and 2 State Elementary School. While the sample was first grade Jatilawang 1 State Elementary School as the experimental class and Kemantran 1 State Elementary School as the control class. Data collection techniques in this study were documentation and tests. The data analysis used in this research (1) mathematical problem solving ability completeness test (one sample t test) and classical completeness test, (2) mathematical problem solving ability average difference test (independent t test), and two proportion difference test, (3) mathematical problem solving ability improvement test. The results of this study are (1) The problem solving ability of students with the RME model assisted by teaching materials exceeds 70, and the proportion of students with the RME model assisted by teaching materials reaches the completion of problem solving ability exceeds 80%. (2) The problem solving ability of students with RME model assisted by teaching materials is better than the problem solving ability of students with conventional learning, and the proportion of students with RME model assisted by teaching materials achieving completeness of problem solving ability is more than or equal to students with conventional learning, (3) the improvement of students' problem solving ability with RME assisted by teaching materials is better than the improvement of students' problem solving ability with conventional learning, (4) the influence of curiousity on problem solving ability is 90%. The implication of this research is that the use of the RME model assisted by Jarimatics can develop problem solving abilities, problem solving abilities with the RME model assisted by Jarimatics are better than conventional models.