An experimental study to analyze descriptively and inferentially the differences in students’ critical thinking ability to learn physics between problem-based learning and conventional model at class XI of SMA Negeri 2 Konawe was performed. Tests and questionnaires were applied. Students of XI IPA classes were served as samples. The samples were determined through a simple random sampling technique, resulting in class XI IPA3 with 31 students as an experimental group, and class XI IPA1 with 31 students as a control group, thereby 62 out of 119 students were selected as the samples. Data were collected by administering a pre- and post-test instrument to measure students’ critical thinking ability to learn Physics. The data were analyzed by using descriptive and inferential statistics. The hypotheses were tested by using the U Mann-Whitney test and Manova analysis, in which a significance level 0.05 was used, thereby if the significance value was lower than 0.05 (sig. < α = 0.05), then H0 was rejected and H1 accepted. The results of data analysis showed that (1) there was a difference in terms of critical thinking ability between the class XI IPA students of SMA Negeri 2 Konawe which learned through a problem-based learning and those who learned in a conventional model and (2) there was a difference in terms of critical thinking ability to learn Physics between the class XI IPA students of SMA Negeri 2 Konawe which learned through a problem-based learning and those who learned in a conventional model.