The purpose of this study is to show the differences in problem-solving ability between first-year University students who received culture-based contextual learning and conventional learning. This research is a quantitative research using quasi-experimental research design. Samples were the First-year students of mathematics education department; Nusa Cendana University consists of 58 students who were divided into two groups each of 29 students. The results showed there are differences in the n-gain average of problem-solving ability significantly between students who receive culture-based contextual learning and conventional learning. The n-gain average of experiment group is 0.51 or medium category while the average n-gain of the control group is 0.29 or low category. Student categories of SNMPTN and Mandiri are significantly different whereas students? category of SBMPTN between the two groups does not differ significantly.