Leadership is regarded as an important factor affecting employee attitudes, behavior and work performance. It has been a hot topic in the fields of organizational behavior and human resource management. Unfortunately, the concept of leadership may have overestimated the effectiveness of leaders. It is not accurate to determine the impact of leadership behavior on subordinates without considering its boundary conditions. Kerr and Jermier （1978） proposed the substitutes for leadership theory, claiming that subordinate individual characteristics, task characteristics and organizational factors can provide guidance and positive effects for employees, influencing the effectiveness of leadership. Evolving from path-goal theory, substitutes for leadership theory becomes a framework to fully interpret the contingent relationship between leadership behavior and outcome variables. It emphasizes more the importance of situational factors compared with former leadership contingent theory. During the several decades after the theory was proposed, the empirical evidence using this theory were inconsistent. What needs to be stressed and explained is that previous measurements of alternative factors are not rigorous, the chosen sample has small variance on some alternative variables, most studies are cross-sectional studies, the methods and procedures of statistical analysis are not unified, and the theoretical model is too complex and huge. Therefore, we cannot overthrow the theory because of the defects of previous empirical research. Entering the new century, with organization flattening and the awakening of individual self-consciousness, individuals and teams cannot be fully motivated only by formal leadership. Although substitutes for leadership theory attracts more and more attention from scholars, there are still many studies that do not grasp the core of the theory when applying it. Responding to the above problems, we systematically review the theory development, its differences from other leadership theories, and core ideas and summarize the flaws of the early empirical studies in order to have more clear understanding of substitutes for leadership. Specifically, substitutes for leadership theory differs from other theories as follows: other leadership theories posit that some leadership styles are universally effective and this theory assumes that no leadership style is universally effective; other leadership theories focus on leader factors and this theory emphasizes the factors unrelated to leaders; different from path-goal theory, this theory elaborates the context factors that leader behavior does not need or is useless in details. We further summarize the central points of this theory, including four statements. Firstly, substitutes for leadership can directly influence the desirable outcomes. Secondly, substitutes for leadership can play a variety of roles, like a moderator or a mediator. Thirdly, for different leader behaviors, the same substitution can exert different influences. Lastly, it is necessary to further distinguish between direct and indirect effects of leader behavior. In addition, we sum up new factors of substitutes for leadership, besides these factors originally proposed by Kerr and Jermier （1986）. From the view of individual characteristics, subordinates’ ability, knowledge, traits, values, cognition and evaluation are explored by researchers as substitutes for leadership. Organizational institutions and culture can also affect the effectiveness of leader behavior, such as human resource management practice, organizational justice, and team shared cognition. To further deepen and develop substitutes for leadership theory, future research needs to expand leadership level, elaborate the roles substitute factors play, construct a new framework, and explore the interactive effect among different substitute factors. We contribute to the literature of substitutes for leadership theory by systematically combing the theory development. The four key pints of the theory we summarize can help researchers to better understand and apply this theory to relevant research questions. Following the classification framework by Kerr and Jermier, we sum up many new substitutional factors, and to a certain extent, enrich the substitutes for leadership theory. The research directions we propose have some implications for the development of the theory.