The paper is the first study to combine four-stage approach and dynamic data envelopment analysis (DEA) to investigate the efficiencies of 45 international tourist hotels in Taiwan during 2002–2011. Using the four-stage approach of Fried et al. [1999. Incorporating the operating environment into a nonparametric measure of technical efficiency. Journal of Productivity Analysis, 12(3), 249–267] and the dynamic DEA model by Tone and Tsutsui [2010. Dynamic DEA: A slacks-based measure approach. Omega, 38(3), 145–156], the study considers the effects of the external operating environments and carry-over activities between two consecutive terms, and calculates managerial efficiency of international tourism hotels in Taiwan. The empirical results show that dynamic DEA provides more stable pure managerial efficiencies than traditional DEA, reflecting the continuity of the links between terms. Then the study employs Tobit regression to estimate the effects of location, management style, and occupancy rate on input slacks. The location in resort area leads to less slacks of employee and room. The chain system helps reduce slacks in room and area of meal department. The slack of employee and room gets lower while occupancy rate gets higher. After controlling the three external environment variables, the efficiency scores of 28 hotels increase, 9 hotels fall, and other 8 hotels still sustain on the efficient frontier.