A comparison between three magnetic resonance coronary angiography (MRCA) respiratory motion suppression techniques was performed for both normal subjects and patients with coronary artery disease (CAD). MRCA images were acquired in 17 normal subjects and 15 patients with CAD, using conventional breath-hold MRCA, navigator echo (NE)-guided breath-hold MRCA (LED feedback), and NE-gated MRCA during free respiration. Image quality, diaphragm registration, and total acquisition time were assessed. Overall, there was poor diaphragm registration for conventional breath-holding compared with free respiration (P < 0.001). CAD patients found it significantly more difficult to perform a steady breath-hold (P = 0.04) or attain the same diaphragm position over multiple breath-holds than normal subjects (P = 0.02). All normal subjects, but only 3 of the 15 CAD patients, were able to perform the LED feedback technique (P < 0.001). For normal subjects, image quality was similar between the three respiratory suppression techniques (P = 0.3), while for CAD patients there was an improvement in image quality, for images acquired during free respiration (breath-hold vs. free respiration, P < 0.01). There was no significant difference in the total acquisition times between the breath-hold and free respiration techniques (P = 0.2). There were substantial differences in the effectiveness of MRCA respiratory suppression techniques between normal subjects and CAD patients. In patients, only NE-gated MRCA performed well, requiring minimal cooperation with no increase in total acquisition time. Validation of NE-MRCA techniques should always be performed in patients, as well as normal subjects, to ensure correct evaluation of the technique for the target population.