This research was motivated by the low mathematical literacy abilities of students. Students are less able to understand, reason, model, formulate and interpret problems in finding solutions, but directly operate them. One of the reasons for this is the inappropriate implementation of learning. The aim of this research is to determine the improvement in students' mathematical literacy skills who use the Problem Based Instruction (PBI) learning model better than the conventional learning model. The type of research used was quasi-experimental with a Pretest-Posttest Control Group Design research design. The research population was class VII students at MTsN 7 Agam. Sampling was carried out using a simple random sampling technique, where class VII.2 was the experimental class and class VII.3 was the control class. Data was collected through a descriptive test and analyzed using the t-test. Obtained t_(count) (3.502)>t_tabel (1.673) means that the increase in mathematical literacy skills of students who use the Problem Based Instruction (PBI) learning model is better compared to conventional learning models.