Autistic children generally have complex obstacles in terms of communication, social interaction and behavior. One of the obstacles in behavior experienced by autistic children is the lack of ability to sort out appropriate and inappropriate behavior to do, that it can trigger impulsive behavior. Impulsive behaviors, in this study, means not able to sit down quietly for a long time, shout if the class feels quite quiet, and do the task in a hurry without paying attention to the commands. Therefore, it is necessary to give a treatment to reduce this behavior, namely the token economy technique. The purpose of this study is to find out the use of token economy technique to reduce impulsive behavior of a student with low autism level in SMP Negeri 14 Banjarbaru. This study used quantitative approach with the SSR method (Single Subject Study), using the design A1-B-A2, where A1 is the baseline 1, B is the intervention and A2 is the baseline 2. The subject in this study is a student in class VIIIA who has low autistic obstacles. The data was collected by doing direct observation and documentation of this study by recording events (frequency). Data was analyzed using graphical visual analysis technique (Visual Analysis of Graph Data) by plotting the data into the graph, and then the data was analyzed by components in each A1-B-A2 condition. The results showed a decrease in impulsive behavior of the student in SMP Negeri 14 Banjarbaru after applying token economy technique. The results obtained at baseline 1 (A1) are 14 times the occurrence of impulsive behavior, intervention (B) is 10 times which shows the impulsive behavior of children begins to decline and at baseline 2 (A2) is 9 times. The estimated trend tends to improve, decreases mean level from baseline 1 (A1) to baseline 2 (A2). Keywords—token economy technique; impulsive behavior; children with low autism level