Currently, the government is aggressively optimizing students' literacy skills, including mathematical literacy skills. Today's problem is that the dominant students still have low mathematical literacy skills. Therefore, one solution is to apply a Problem Based Learning (PBL) model based on portfolio assessment. This study aims to determine the difference between students' mathematical literacy skills taught using the PBL learning model about portfolios and students' mathematical literacy skills using the direct learning model. Quasi-experimentation is the type of research using the Posttest Only Control Group Design. The population of the study was the seventh-grade students of SMPN 2 Cibal, who found 96 people spread over 3 classes. Samples were taken from the population in a randomized manner so that 32 students of class VIIC were obtained as the experimental group and 32 students of class VIIA as the control group. Data was collected from description testing and also equipped with student portfolio assessment results. The results of data analysis showed that the literacy skills of students who were taught using the PBL learning model showed a better portfolio than the literacy skills of students who used the direct learning model.