Does oil wealth promote or inhibit prospects for civil war? Empirical evidence relating oil to civil war onset is mixed, and depends on the aims of the rebellion: although separatist civil wars (in which rebels aim to create an autonomous region or independent state) occur more frequently in oil-rich regions, oil-rich countries experience fewer center-seeking civil wars (in which rebels aim to capture the capital city). This article provides a new theoretical framework in which the challenger strategically chooses its civil war aims. I first incorporate strategic civil war aims into a formal bargaining model with commitment problems. Then, I derive two countervailing theoretical effects of economic activities, such as oil production, that provide an easy source of government revenues: a conflict-suppressing revenue effect (more money for the government) and a conflict-enhancing predation effect (more for the rebels to capture). Finally, I highlight two reasons that the magnitude of the oil predation effect is larger for separatist than for center-seeking challengers, which connects the theoretical implications to the motivating empirical pattern. First, a strategic selection effect for ethnic minorities: governments face more severe commitment problems toward small ethnic groups — who prefer separatist over center-seeking civil war. Second, a geography of rebellion effect: oil-funded repression more effectively deters center-seeking challenges than peripheral insurgencies.