Oshiboris, wet hand towels offered to customers in places such as restaurants or bars, are used to clean one’s hands before eating. Long part of hospitality culture in Japan, many restaurants and bars do not prepare them in-house but employ a rental service for laundering, shaping, and delivering them already damp. There are numerous rental service companies offering oshibori rental services in Japan. A crucial task for these companies is controlling the total number of oshiboris. They lend oshiboris to customers, such as restaurants and bars, and collect the used ones afterward. However, the number of collected oshiboris often falls short of the quantity lent. This discrepancy causes significant losses for oshibori rental service companies, making it vital to increase the return rate of oshiboris. The rental system involves charging a fine for unreturned oshiboris. However, since the oshibori rental company is unable to track the number of unreturned oshibori accurately, it has become difficult to charge for them. If the company can track the number of returned oshibori with sufficient precision, it would be possible to charge accordingly. Furthermore, if customers begin to avoid this situation, the number of returned oshiboris will likely to increase. It’s practically infeasible to count the returned oshiboris at pickup site due to the time constraints faced by delivery staff during collection. Thus, only an approximate guess is possible at that stage. In this study, we propose a method to estimate the return rate of customers using data from actual oshibori rental services. An accurate estimate would enable rental service companies to take necessary actions with their customers, potentially reducing oshibori losses.