Background —The effect of long-term smoking on coronary vasomotion and vasodilator capacity in healthy smokers is unknown. Methods and Results —Myocardial blood flow (MBF) was quantified with [ ¹³ N]ammonia and positron emission tomography (PET) at rest, during cold pressor testing (endothelium-dependent vasomotion), and during dipyridamole-induced hyperemia in 16 long-term smokers and 17 nonsmokers. MBF at rest did not differ between the 2 groups. Cold induced similar increases in rate-pressure product (RPP) in smokers and nonsmokers. However, MBF increased only in nonsmokers and was, during cold, higher than in smokers (0.91±0.18 versus 0.78±0.14 mL · g ⁻¹ · min ⁻¹ , P <0.05). MBF normalized to the RPP (derived from the ratio of MBF ([milliliters per gram per minute] to RPP [beats per minute times millimeters of mercury] times 10 000) declined in smokers but remained unchanged in nonsmokers (0.86±0.10 versus 0.72±0.11, P =0.0006, and 0.99±0.25 versus 0.96±0.27, P =NS). The hyperemic response to dipyridamole and the myocardial flow reserve did not differ between the 2 groups. In a multiple regression model adjusted for age, sex, serum lipid levels, years of smoking, and pack-years, years of smoking was the strongest predictor of the normalized blood flow response to cold ( P <0.001), followed by the HDL/LDL ratio. Conclusions —The normal hyperemic response to dipyridamole in long-term smokers indicates a preserved endothelium-independent coronary vascular smooth muscle relaxation, whereas the abnormal response to cold suggests a defect in coronary vasomotion likely located at the level of the coronary endothelium. Its severity depends on the total exposure time to smoking.