This study aims to determine the effect of the use of contextual teaching  and learning learning models on students' mathematical problem solving abilities and student learning independence compared to conventional learning models. This type of research is a quasi-experiment with a research design, namely the nonequivalent pretest-posttest control group design. The population in this study was all grade VII students of SMP Negeri 1 Tirtayasa. Samples were taken using purposive sampling  techniques so that the samples obtained were class VII-E students as a control class and class VII-F as an experimental class with 38 students in each class. The data collection technique uses 5 items of essay tests, students' mathematical problem-solving abilities, and learning independence questionnaires. The data were analyzed by data normality test, data homogeneity, t-test, as well as quantitative descriptive analysis. Results were obtained (1) independent test sample t-test (posttest) in the experimental class and control class respectively obtained a calculated t-value > ttable so that H0 was accepted and H1 was rejected. (2) Independent sample t-test  (posttest) test in experimental class and control class obtained tcalculate value < ttable so that H0 was rejected and H1 was accepted. Moreover, (3) the results of the analysis of student learning independence were obtained on average by 75% with solid criteria.