The mathematical problem solving ability of students at Universitas Islam Negeri Sjech Djamil Jambek Bukittingi is still relatively low. The flipped classroom learning model assisted by YouTube is one of the solutions. The purpose of this study was to determine the difference in mathematical problem solving ability of students who learn with the flipped classroom learning model assisted by YouTube and the mathematical problem solving ability of students who learn with conventional learning. This research is a type of pseudo-experimental research with a randomized control group only design. The population of this study were students of the Islamic economics study program at the State Islamic University Sjech M. Djamil Djambek Bukit Tinggi with Islamic economics class C as the experimental class and Islamic economics D as the control class. The number of students from each class is 30 people. The instrument used is a test of mathematical problem solving ability. From the research results, the average mathematical problem solving ability of students in the experimental class was 78.47 while the control class was 68.73. The analysis results of the classical assumption test showed that the students' mathematical problem solving ability was normally distributed and had a homogeneous variance. Testing with independent sample t-test results in the value of t count (3.130) t table (2.002) and sig value (0.003) 0.05. The results of this test explain that there is a significant difference in mathematical problem solving ability in the experimental class using the flipped classroom learning method assisted by YouTube with the control class using conventional learning. Based on the findings of this study, it is recommended for lecturers to have their own YouTube channel that helps them to deliver lecture material, so that it can be accessed and used at any time by students.