Purpose: IMRT and partial breast irradiation (PBI) are used to improve cosmetic outcomes in early stage breast cancer patients; however, homogeneity and dose coverage are degraded by respiratory motion during delivery. We investigate the impact of respiratory motion on PBI IMRT using population respiratory data. Methods: The fluences from five different PBI IMRT plans were convolved with a population probability density function (PDF) to simulate dose delivery during respiration. Plan quality and coverage of the CTVsu (CTV plus 5 mm set‐up margin, trimmed back from skin and chestwall) were evaluated using DVH, hotspot, cold spot, and uniformity index. Results: All plans showed blurring (loss of coverage) in the shoulder region of the DVH curves for CTVsu and increased hotspot to the ipsilateral breast. For the patient with the largest breast volume and the smallest DEV:breast volume ratio, the CTVsu is surrounded by tissue. In this patient, the CTVsu coverage is slightly degraded and the hot spot becomes slightly hotter. The patients with a target volume near skin, lung or severe contour changes experience coverage loss near inhomogeneities and increased hot spots. This occurs for trimmed DEVs, although flash was added appropriately to all plans and coverage to the PTV (not only DEV) is adequate in the original plan. In patients with small breasts the CTVsu was noticeably hotter under respiratory conditions than static with the hot spot (2cc) in the CTVsu increasing by 7% and the 95% coverage increasing by 2%. Conclusions: The impact of respiratory motion on PBI is nontrivial and mainly patient dependent. Breast and target shape are the major determining factors in degradation of plan quality with respiration.