This article aims at finding out (1) the differences in the mathematics learning outcomes of the students who follow NHT type cooperative learning with the students who follow direct learning, (2) the influence of interaction between the learning model used and the students’ learning interests, (3) the difference in the mathematics learning outcomes between the students who follow NHT type cooperative learning and the students who follow direct learning reviewed from high learning interests, (4) differences in students’ mathematics learning outcomes between the students who follow NHT type cooperative learning and the students who follow direct learning reviewed from low learning interests. This the type of pseudo-experimental research. The instruments used are test of learning outcomes, questionnaires of students' learning interests, and observation of the Lesson Plan implementation. The analytical techniques used are descriptive analysis techniques and inferential statistical analysis with ANOVA tests, and Tukey advanced tests. Based on the results of the ANOVA test and Tukey advanced test obtained that there are significant differences in the mathematics learning outcomes between using a cooperative learning model NHT type and the direct learning model in the students who have a high interest in learning. Based on the data analysis, it is obtained the significant value of < α (0.010 < 0.05) then H0 is rejected. Therefore, it is concluded that there are significant differences in mathematics learning outcomes between using NHT-type cooperative learning model and the direct learning model in the students who have a low interest in learning.