This research is aimed to attest whether traditional game with experiential learning method can improve social competence of school age children and whether this improvement can be moderated by gender. This research is a quasi-experiment with the design of multiple treatment and control with pretest which involves three subject groups. Experiment-I Group plays traditional game with experiential learning method. Experiment-II Group plays traditional game without experiential learning method. Control group is without treatment. The subject of research is the fifth-grade of Elementary School which the group is differentiated between male and female totaled to 168 students. The measuring tool is using two equal scales of social competence. Scale I is used to measure pretest whereas Scale II is used to measure posttest. Data analysis is using Two-Ways Multivariate ANAVA to distinguish gain score difference between Experiment-I Group, Experiment-II Group, and Control. Result of research indicates that there is an effect of traditional game with experiential learning method or BERLIAN (Bermain-ExpeRiential Learning-ANak) and without BERLIAN on the improvement of child social competence (F = 19.915; p = 0.000). Traditional game with BERLIAN method can improve social competence compared to traditional game without BERLIAN method. Indeed, traditional game without BERLIAN method can improve the aspects of problem-solving and cooperation, but traditional game with BERLIAN method may increase whole aspects of social competence, such as problem-solving, self-control, cooperation and empathy. This research has explained that different role of gender is not quite influential to the improvement of social competence.