This research was based on the background of the lack of critical thinking skills of students in SMP PGRI 6 Bandar Lampung. This is caused by several factors, one of which is the selection of learning models such as conventional learning models that are continuously applied by teachers in the classroom. Efforts to address these problems, researchers aim to determine whether there is the effect of problem-based learning with cognitive conflict strategies towards students' critical thinking skills better than conventional learning. This research is a type of experimental research from Quasi Experimental design. This study uses the Pretest design - Posttest Control Group. The subjects in this study were VII grade students of PGRI 6 SMP Bandar Lampung with a total sample of 62 students. The sampling technique was carried out by means of class random techniques. Instruments used to collect data in tests of critical thinking skills. For the normality test using the Liliefors test, the homogeneity test uses a similarity test of two variances. The hypothesis test used in this study is the t-test on n-gain data. With a sample of 62 students and a significant level of α = 0.005 obtained t table = 2,0002 and the calculation results obtained by tcount = 2.50791, it can be seen that tcount> ttable then H0 is rejected. Thus it can be concluded that whether there is the influence of problem-based learning with cognitive conflict strategies towards students' critical thinking skills better than conventional learning.