Cell phones with GPS function as well as GPS loggers are widely used and we can easily obtain users' geographic information. However, still battery consumption in these mobile devices is main concern and then we are not allowed to obtain GPS positioning data so frequently. In this paper, we propose a stayed location estimation method for sparse GPS positioning data. After generating initial clusters from a sequence of measured positions, we set the effective radius for every cluster based on positioning accuracy and merge the clusters effectively using it. After that, we temporarily remove short-time clusters but do not remove measured positions included in them. Then we merge the clusters again, taking all the measured positions into consideration. We perform this process twice, i.e, we perform two-stage short-time cluster removal, and finally realize accurate stayed location estimation even when the GPS positioning interval is five minutes or more. Experiments demonstrate that the total distance error between the estimated stayed location and the true stayed location is reduced by more than 50% compared to a conventional state-of-the-art method.