The purpose of this study was to improve student learning outcomes in seventh grade junior high school students by using a contextual approach to social arithmetic material. The method used in this research is Classroom Action Research (CAR) which is divided into 3 (three) cycles. The population in this study were students of class VII-B SMP PGRI 5 Cimahi with a total of 31 students consisting of 17 male students and 14 female students. The data collection process is carried out in a planned and systematic manner to monitor the process of learning activities, increase student motivation and improve learning outcomes. Data obtained through questionnaires and cognitive ability tests. The data processing technique in this research is by calculating the percentage of the average of all students and then comparing it with the KKM score at the school. With a contextual approach to social arithmetic material, there has been a significant increase in student learning outcomes from cycle I to cycle III. Completeness of learning in cycle I was 54%, which means that in cycle I there were 17 students who had been able to reach KKM. In the second cycle the completeness increased to 77%, meaning that there were 24 students who had reached the KKM. In cycle III completeness increases to 100%, meaning that in cycle III all students can achieve the KKM score, the achievement of student learning outcomes that meet classical learning completeness based on the average score appears to have met the completeness criteria indicated by the acquisition percentage of 100%. The average score of the first cycle was 67.09, increasing in the second cycle to 74.67, and increasing in the third cycle to 87.58. The highest value in cycle I was 85, increased in cycle II to 90 and increased again in cycle III to 100. The lowest value of cycle I was 50, increased in cycle II to 60, and increased again in cycle III to 75. The meaning of individual completeness was each students have been able to achieve the KKM of 75. Achievement of student learning outcomes that meet classical learning completeness based on the average score appears to have met the completeness criteria indicated by the acquisition percentage of 100%.