This study aims to determine the effect of students' understanding of Mathematics learning using the discussion method learning model with students using conventional learning models. This study uses a quasi-experimental design with Nonequivalent Control Group Design. The population subjects in this study were all students of class IV SDN Doyong 04 Tangerang City, totaling 52 students, by taking samples of two classes, namely class IV A which amounted to 23 students as the control class and class IV B as the experimental class. Determination of the sample is done using Nonprobability Sampling. The data collection technique used an essay test instrument consisting of 10 valid and reliable questions. For testing the pretest hypothesis in the study   = 5.232 and   = 2.073. while for testing the experimental class from the t-test results obtained   = 8,339 and   = 2,048 at the significance level (a = 0.05), it can be concluded that it is   rejected and    accepted, meaning that there is an effect of using the discussion method on the learning outcomes of fourth grade students, which is significant between the average value of the experimental class and the control class. This can be interpreted that the learning outcomes of students in Mathematics learning FPB using the discussion method are more effective than using conventional methods.