21st-century education requires students to master mathematical problem-solving skills and self-regulated learning (SRL), but these skills remain a challenge for junior high school students in Indonesia. This study aims to examine the impact of implementing a Collaborative Learning Model of the PBL combined with a scientific approach (observing, questioning, reasoning/trying, communicating) on students' mathematical problem-solving skills and SRL. The study used a pre-experimental method with a One-Group Pretest–Posttest design on 31 seventh-grade students at a junior high school in Sleman Regency. Data on problem-solving skills were collected through a 5-item essay test with a maximum theoretical score of 50, while SRL was measured using a Likert scale questionnaire. Inferential analysis used the Wilcoxon signed-rank test for problem-solving skills (the assumption of normality was not met) and the paired sample t-test for SRL. The results showed a significant increase in problem-solving skills from 15.32 to 30.92, with posttest achievement equivalent to 61.84% of the theoretical maximum score. SRL also increased significantly from 121.81 to 122.74, although the mean difference was relatively small compared to the scale range. These findings indicate changes in students’ achievement after the intervention; however, the results should be interpreted cautiously due to the short intervention duration and the limitations of the one-group pretest–posttest design without a control group.