This study aims to determine the average difference in physics problem solving ability between students who use cooperative learning model type think-pair-share (TPS) with students who use direct instruction (DI) learning model on the concept of dynamic fluid. The population in this study were all students of class XI MAN 2 Probolinggo, the research samples used were class XI IPA 1 and XI IPA 5 which were selected through simple random sampling. The research instruments used were 25 multiple choice test questions and a learning motivation questionnaire consisting of 12 statements, both of these instruments had gone through the instrument eligibilityt test. The prerequisite test for data analysis was carried out through the normality test and homogeneity test, the data on students' problem solving ability’ results of the normality test with Shapiro-Wilk in the four class groups were normally distributed, the test continued with the homogeneity test through Levene's Test the result was the data on problem solving ability between learning models and between groups was homogeneous. The results of hypothesis testing with Two-Way Anava are (1) there is a difference in physics problem solving ability on the concept of dynamic fluid between students who use cooperative learning model type think-pair-share (TPS) with students who use direct instruction learning model (DI), (2) there is a difference in physics problem solving ability on the concept of dynamic fluid between students who have high learning motivation and students who have low learning motivation, and (3) there is an interaction effect between the cooperative learning model type think-pair-share (TPS) with student learning motivation on problem solving ability on the concept of dynamic fluid.