This study aims to determine the effect of the Numbered Head Together (NHT) and Expository Learning Model (ELM) learning models on student learning outcomes in derived mathematics. The method used in this research is a quasi-experimental method, namely the Nonequaivalent Control Group Design. The experimental class was given a Numbered Head Together (NHT) The NHT model for class XI SMA IPA 5 and XI SMA IPA 1 as conventional classes is given in the form of the ELM model. Data analysis used the t test, where the results showed: 1) there was a significant increase in student learning outcomes in derived material using the Learning Numbered Head Together (NHT) model, 2) student learning outcomes in mathematics were higher using the Learning Numbered Head Together model. (NHT) than the learning outcomes obtained with the Expository Learning Model (ELM) model. This can be seen from the test results with the t. Test, mark Tcount = 89,12 And value ofT’ table = 1,67 so that TCount>T’table = 89,2>1,67, so rejected and accepted. In conclusion, students who are taught mathematics using the NHT model understand much more and have very good results compared to the ELM model. Suggestions, in implementing mathematics material, it is expected that educators use the NTH learning model to improve learning outcomes and produce good thinking concepts in mathematics.