This research was an experimental study aimed at collecting information about increasing Mathematical comprehension and problem solving skills of vocational students through a REACT learning strategy. The population in this research was the X grade students of a Vocational School in Jakarta, and the sample consisted of two grades. The experimental class was a class that learned trigonometry using REACT strategy. While another class, called control class, was a class treated trigonometri teaching by using conventional strategy. The instrument used to collect data was a comprehension and problem solving test, in essay type and observation sheet. The data of test results were the results of pretest and post test in the form of normalized gain (N-Gain then analyzed quantitatively). Statistical tests used were t-test for N-Gain score of Mathematical comprehension and Mann-Witney test for N-gain score of Mathematical problem solving, using SPSS 16.0. The analysis of test results obtained with the t-test Asym.Sig. (1-tailed) = 0.000 and the Mann-Witney obtained Asym.Sig. (1-tailed) = 0.000. The observation data was analyzed qualitatively and the result showed that learning with REACT strategy was run well and effectively built a condusive class for learning. When students were able to find or discover the knowledge, it would build their self confidence to find other knowledge through. The result showed that (1) increasing mathematics comprehension with REACT was higher/ better than the conventional learning. (2) increasing students' mathematics problem solving skills with REACT was higher/ better than students in conventional learning.