The purpose of this study was to determine the effect of learning models and nutritional status on the physical fitness of grade VII students of SMPN 21 Padang. This research model is a pseudo-experiment designed by Factorial by Level 2x2. The population of this study amounted to 119 people and a sample of 60 people with the Purposive Random Sampling technique. Collection of Nutritional Status data by tests using Body Mass Index (BMI) and Physical Fitness using the Indonesian Physical Fitness Test (TKJI) Age 13-15 years. Data were analyzed using two-path ANAVA at significance level α=0.05. Furthermore, because no interaction was found between the learning model and physical fitness, the Tukey Test was not carried out. The results of the data analysis showed: (1) Overall there was no difference in the learning outcomes of the PjBL learning model with the Case Method learning model, (2) There was no interaction between the learning model and nutritional status on physical fitness, (3) Physical fitness of students treatment of PjBL learning model is better than Case Method learning model in high nutritional status group, (4) Physical fitness of students treatment of PjBL learning model is better than Case Method learning model in low motor group.