In breast cancer treatment, intensity-modulated radiation therapy (IMRT) can be utilized to deliver more homogeneous dose to target tissues to minimize the cosmetic impact. We have investigated the effect of the respiratory cycle and radiation beam-on timing on the dose distribution in free-breathing dynamic breast IMRT treatment. Six patients with early stage cancer of the left breast were included in this study. A helical computed tomography (CT) scan was acquired for treatment planning. A four-dimensional computed tomography (4D CT) scan was obtained right after the helical CT scan with little or no setup uncertainty to simulate patient respiratory motion. After optimizing based on the helical CT scan, the sliding-window dynamic multileaf collimator (DMLC) leaf sequence was segmented into multiple sections that corresponded to various respiratory phases per respiratory cycle and radiation beam-on timing. The segmented DMLC leaf sections were grouped according to respiratory phases and superimposed over the radiation fields of corresponding 4D CT image set. Dose calculation was then performed for each phase of the 4D CT scan. The total dose distribution was computed by accumulating the contribution of dose from each phase to every voxel in the region of interest. This was tracked by a deformable registration program throughout all of the respiratory phases of the 4D CT scan. A dose heterogeneity index, defined as the ratio between (D20-D80) and the prescription dose, was introduced to numerically illustrate the impact of respiratory motion on the dose distribution of treatment volume. A respiratory cycle range of 4-8 s and randomly distributed beam-on timing were assigned to simulate the patient respiratory motion during the free-breathing treatment. The results showed that the respiratory cycle period and radiation beam-on timing presented limited impact on the target dose coverage and slightly increased the target dose heterogeneity. This motion impact tended to increase the variation of target dose coverage and heterogeneity between treatment fractions with different radiation beam-on timing. The target dose coverage and heterogeneity were more susceptible to the radiation beam-on timing for patients with long respiratory cycle (longer than 6 s) and large breast motion amplitudes (larger than 0.7 cm). The same results could be found for respiratory cycle up to 8 s and respiratory motion amplitude up to 1 cm. The heart dose distribution did not change significantly regardless of respiratory cycle and radiation beam-on timing.