This article examines the basis of military leadership preferences toward war termination, and thus the basis of the associated civil–military divide, within the context of protracted small war. The conventional wisdom posits military leadership preferences as a near-constant in favor of persistence and thus expects a dominant pattern of military obstructionism. However, such a pattern does not hold empirically across the population of small wars. This gap between expectation and evidence derives at least in part from the limits of the bureaucratic–organizational model, focused on the military’s desire for resources, autonomy, and influence, that underlies the conventional wisdom. In contrast, this article suggests an alternative model privileging the demands of institutional legitimacy. The legitimacy motive as conceptualized here is particularly salient to the small war context. It accordingly provides a foundation for better understanding variation in military leadership preferences toward war termination and thus variation in the direction and intensity of the civil–military divide.