The research investigated the rules relevant with leaving reasons and personal characteristic variables of nursing staff by rough set theory (RST). RST can discover the important facts hidden in the data and describe them with natural language. That is it can describe the information provided by researchers in a broad sense and delete the redundant part of original data for the decision rules setting. It is expected the research can establish rules related with leaving reasons and personal characteristics variables of nursing staff. The study conducted analysis on 458 persons' data that include six attributes - age, marriage status, education, seniority, service department and dimission reasons to propose concerned suggestions and comments based on the research result as reference for Hospital A and other hospitals as well as relevant academic institutions.