This study compares Learning Start with a Question (LSQ) with traditional learning to investigate how to improve mathematical critical thinking skills. Quantitative research methods are used in a quasi-experimental way. The sample of this study was 83 junior high school students in Ternate City, Indonesia. An essay test was used, with questions based on mathematical critical thinking skills. Inferential statistics and an independent sample t-test were used to analyze the data. Based on the study's results, with the application of LSQ learning, most students can solve problems in critical mathematical thinking correctly, assisted by steps and indicators of completeness of mathematical critical thinking skills. Students who use LSQ learning have a higher average test score than traditional learning students. The results of the t-test calculation show that students who learn to use LSQ learning have differences in increasing mathematical critical thinking skills with students who follow traditional learning.