Many studies already indicated that cooperative learning can promote learners' interest than learning process purely based on individual competition, so that learners could get more enthusiastic while they can get involved in learning activities. Nevertheless, cooperative learning usually groups students by means of random or normal distribution approach, it is not a thoughtful means for learners find proper cooperative partners, and it causes poor interaction among learners in the same group. If the learners from the same group are not properly working with each other, it could not fit the ultimate goal of cooperative learning. Therefore, this study attempts to utilize wireless sensing technology as a tool to detect learner's interaction by sensing two or more learners' distances in certain amount of time. Based on the mining social interaction, this study claims that social interaction should be considered as an essential factor while formulating cooperative learning groups. In hope, the study could help learner achieve better learning performance of cooperative learning than randomly grouping method.