Blended learning is one of the many learning models that can be applied to learning activities at this time. The study was conducted to determine the effect of blended learning on the ability to solve problems in the subject history of Islamic culture. The study used a quasi-experimental quantitative approach (nonequivalent control group design). The data collection technique used a test (essay) given at the pre-test and post-test; the data analysis used a prerequisite test, namely, the normality test and homogeneity test, followed by an independent sample t-test. The research location was at MTsN 1 Jombang, involving 120 students, 60 students as the experimental group applying blended learning and 60 students as the control group applying conventional learning. The study's results based on the independent sample t-test showed that the pre-test scores of the two groups had no significant difference, namely tcount< ttable (0.393 < 2.001) with sig. (2–tailed) 0.695 > 0.05, it can be ascertained that there is no difference between the pre-test data of the experimental and control groups. And the results of the independent sample t-test showed that the post-test scores of the two groups had significant differences, namely the results tcount> ttable (12.370 > 2.001) with sig. (2-tailed) 0.000 < 0.05. So it can be ascertained that there is a difference between the post-test of the experimental and control groups. The conclusion is that blended learning significantly affects problem-solving ability in the subjects History of Islamic Cultural at MTsN 1 Jombang in the experimental group compared to conventional learning in the experimental group by reviewing the post-test results of the two groups. The pre-test of the two groups was conducted to determine students' initial ability before being given treatment.