Students' thinking styles are important factors for teachers to accommodate all types of students' thinking styles in a learning approach. Therefore, this study aims to determine students' mathematical problem-solving skills from their respective thinking styles using a metacognitive approach. The sample in this study was 60 students of class XII Madrasah Aliyah Negeri Ambon consisting of 31 Mathematics IPA-1 classes as the experimental class and 29 Mathematics IPA-2 classes as the control class. This research is a quasi-experimental study with a pretest-posttest control group design. The research instruments used were tests (pre-test and post-test) and the thinking style instrument. The data were tested by univariate test and t-test. The experimental class concluded; There are differences in the improvement of students' mathematical problem-solving abilities between groups of thinking styles and there is an influence of thinking styles on problem-solving abilities. While in the control class; There is no difference in increasing students' mathematical problem-solving skills between groups of thinking styles, nor is there any influence of thinking styles on problem-solving abilities. These results illustrate that students' mathematical problem-solving skills can be influenced by thinking styles as a result of the learning approach used, namely the metacognitive approach.