Power shifts between interstate dyadic relationships are generally considered one of the most important factors affecting the preventive motivation for war. It is also a prominent cause of war in power transition theory. This theory predicts that war is likely to occur when the power of a challenger rises to parity with that of a dominant power. Based on this theory’s prediction, this paper can state two basic hypotheses as follows: (1) If power parity exists, preventive war is more likely; and (2) If power preponderance exists, preventive war is less likely. However, despite the second hypothesis indicating that war is less likely with power preponderance, the United States has often considered and launched preventive wars against middle and small powers. By critically evaluating the relevance of the above hypotheses given preventive wars launched and considered by the U.S. after World War II, this paper argues that the expected cost of war has been a crucial factor affecting the decision to engage in preventive war. In this context, this paper argues that the absence of war between the U.S. and the Soviet Union in the Cold War era was caused largely by the prohibitively high expected cost of war rather than by the preponderant power of the U.S. over the Soviet Union. This argument can also be applicable to the confrontation between the U.S. and other major and strong middle powers like China and North Korea. However, the U.S. launched preventive wars against weak middle and small powers such as Iraq and Grenada largely because the expected costs of wars in these conflicts were very low. In order to verify these arguments, this paper presents the modified version of the hypotheses of power transition theory and identifies intervening variables affecting the U.S. decisions in favor of preventive war.