This study has the aim of knowing the development of problem solving abilities before and after the implementation of each model and knowing whether or not there are differences in problem solving abilities between the models in the experimental class 1 and experimental class 2. The approach used in this study is quantitative with a quasi design. experimental. The sampling technique was purposive sampling and class X IPS 2 was taken as the experimental class 1 which was given the PBL model with index card match. Meanwhile, class X IPS 3 as experimental class 2 was given a PBL model with think pair share. The results of the analysis in this study obtained conclusions including: 1) experimental class 1 there was a significant development of problem solving ability in the results before and after giving the model, 2) experimental class 2 there was a significant development of problem solving ability in the results before and after giving the model, 3) there is no significant difference in problem solving ability between experimental class 1 and experiment 2. Thus it can be concluded that the two models are equally able to improve students' problem solving abilities.