ABSTRACT   This study aims to describe the effectiveness of students’ mathematical problem solving ability through Think Pair Share (TPS) learning model at the tenth grade students of SMK Negeri 1 Sarudik. The research was conducted by using quasi experimental method (pretest post test one group design) with 25 students as the sample and they were taken by using cluster random sampling technique from 105 students. Test and questionnaire were used in collecting the data. Based on descriptive analysis, it could be found (a) the average of using TPS learning model was 3.39 (very good category) and (b) the average of students’ mathematical problem solving ability before using TPS learning model was 54,53 (fair category) and after using TPS learning model was 85.2 (very good category). Furthermore, based on inferential statistic by using paired sample ttest   and helping SPSS version 16, the result showed the significant value was less than 0.05 (0.000<0.05). It means, there is a significant difference of students’ mathematical problem solving ability before and after using TPS learning model. In the other words, TPS learning model was effective was used to increase students’ mathematical problem solving ability.   Keywords: Think Pair Share (TPS) learning model, students’ mathematical problem solving ability