This study was performed to assess the role of additional myocardial perfusion imaging during high dose dobutamine/atropine stress magnetic resonance (DSMR-wall motion) for the evaluation of patients with intermediate (50–70%) coronary artery stenosis. Routine DSMR-wall motion was combined with perfusion imaging (DSMR-perfusion) in 174 consecutive patients with chest pain syndromes who were scheduled for a clinically indicated coronary angiography. When defining CAD as the presence of a ≥ 50% stenosis, the addition of perfusion imaging improved sensitivity (90 vs. 79%, P < 0.001) with a non-significant reduction in specificity (85 vs. 90%, P = 0.13) and an improvement in overall diagnostic accuracy (88 vs. 84%, P = 0.008). Adding perfusion imaging improved sensitivity in patients with intermediate stenosis (87 vs. 72%, P = 0.03), but not in patients with severe (≥70%) stenosis (93 vs. 84%, P = 0.06). In patients with severe stenosis specificity of DSMR-perfusion versus DSMR-wall motion decreased (61 vs 70%, P = 0.001) resulting in a lower overall accuracy (71 vs 74%, P = 0.03). Using a cutoff of ≥50% for the definition of CAD, sensitivity of DSMR-perfusion compared to DSMR-wall motion was significantly higher in patients with single vessel (88 vs. 77%, P = 0.03) and multi vessel disease (93 vs. 79%, P = 0.03), whereas no significant differences were found using a cutoff of ≥70% stenosis for the definition of CAD. The addition of perfusion imaging during DSMR-wall motion improved the sensitivity in patients with intermediate coronary artery stenosis. Overall diagnostic accuracy increased only when defining CAD as ≥50% stenosis. In patients with ≥70% stenosis DSMR-wall motion alone had higher accuracy due to more false-positive cases with DSMR-perfusion.