This study aims to determine; (1) Differences in mathematics learning outcomes between learning with reciprocal teaching learning models and facilitator student models and explaining after controlling students' initial abilities; (2) Effect of interaction between learning models and learning independence on mathematics learning outcomes after controlling students' initial abilities; (3) Differences in mathematics learning outcomes between those who used the reciprocal teaching learning model with high learning independence and who studied using student facilitator and explaining model with high learning independence after controlling students' early ability; (4) Differences in mathematics learning outcomes between those who used the reciprocal teaching learning model with low learning independence and who studied using student facilitator and explaining model with low learning independence after controlling students' early ability. The research method used is an experimental method. Data analysis used descriptive analysis and statistical analysis using two-way covariate analysis. The result of hypothesis 1 testing shows that H0 is rejected based on F-Test, line A with value Fc = 4.47 is greater than Ftable (0.05; 1: 59) = 4.00. The result of hypothesis 2 testing shows that H0 is rejected based on F Test statistic, interaction row value Fcount = 14.94 bigger than Ftable (0.01; 1: 59) = 7.08. The results of the hypothesis 3 testing analysis show that H0 is rejected based on the t-test statistic, the value of t count = 4.90. This value is greater than t table (0.01; 59) = 2.39. The results of the analysis of testing hypothesis 4 shows that H0 is rejected based on the t-test statistic, the value of t count = 1.83. This value is greater than t table (0.05; 59) = 1.67.