The purpose of this study was to indentify a variety of factors that affect satisfaction and commitment on close relationship. We attempted to explain the process by which close relationships develop and deteriorate over time, by testing the adequacy of the investment model using longitudinal self-reported data from college students. The result indicated that changes over time in commitment were more strongly predictive of stay/leave behaviors than were other investment model factors. Thus, stay/leave behaviors were mediated by changes over time in commitment. Contrary to the investment model, variations in rewards contributed significantly to the prediction of satisfaction, but costs did not. And sex differences in the variables contributed to commitment was found; whereas men increased commitment because of increases in satisfaction, women promoted because of increases in satisfaction and investment size. These findings suggested that it is necessary to revise the investment model. From this point of view, we discussed the generalizability and limitation of investment model and sex differences in the process of stay/leave behaviors.