<![CDATA[This study aims to describe mathematical problem solving skills using the Two Stay Two Stray type cooperative learning model better than using conventional learning. The type of research used is a quantitative approach to the experimental method by connecting two variables, namely the cooperative learning model type Two Stay Two Stray as the independent variable and mathematical problem solving ability as the dependent variable. Posttes only control design research design. The population of this study were all students of class IX of SMP Negeri 2 Merangin as many as three classes consisting of 72 students. The sampling technique uses simple random sampling. Data collection techniques through tests of mathematical problem solving abilities in the form of essay questions on the material of the Roots and Forms of Roots. The data analysis technique used to test hypotheses is the t-test. From the analysis of the final test data mathematical problem solving abilities in the experimental class students obtained an average value of 34.04, while the control class students obtained an average value of 31.292. Based on the test requirements, the data analysis is normally distributed and homogeneous variance, then using a hypothesis test with the t-test obtained, and. Because it is accepted and rejected. Conclusions from the research on mathematical problem solving abilities using the two stay two stray model are better than the mathematical problem solving abilities using conventional learning of class IX students of SMP N 2 Merangin 2015/2016 ]]>
