This dissertation attempts to uncover systematic patterns regarding elite management in dictatorships. To do so, it describes how dictators manage their elites and what factors determine the outcomes of their decision. Although considerable literature has examined the various structural features of dictatorships and has identified different elite management strategies to explain the persistence of dictatorships, few, if any, have empirically tested any of the theoretical propositions generated by this increasingly large body of literature. This dissertation is the first empirical attempt to explore the elite management strategies of various dictatorships, ranging from the individual case of the most extreme dictatorship (North Korea) as well as different kinds of military dictatorships (South Korea), and global patterns of autocratic regimes. To address the main research question, "what determines the choice of the dictator's elite management strategies?" this dissertation identifies three key factors - dictator, elites, and structure. The relationship between dictators and elites is basically hostile. Conflicts between actors over power acquisition often emerge in violent ways. Nevertheless, dictators do not always treat elites with repression. They sometimes make efforts to embrace and cooperate with the other elites. The variation of their strategies toward elites is determined by various conditions. The results of this dissertation indicate that three factors of dictators, elites, and the environments surrounding them significantly affect the elite management strategies under dictatorships. The main theoretical arguments of this dissertation are supported by the evidence in the three empirical chapters of this dissertation.