The interplay between a mobile target and a dynamic multileaf collimator can compromise the accuracy of intensity-modulated radiation therapy (IMRT). Our goal in this study is to investigate the dosimetric effects caused by the respiratory motion during IMRT. A moving phantom was built to simulate the typical breathing motion. Different sizes of the gating windows were selected for gated deliveries. The residual motions during the beam-on period ranged from 0.5 to 3 cm. An IMRT plan with five treatment fields from different gantry angles were delivered to the moving phantom for three irradiation conditions: Stationary condition, moving with the use of gating system, and moving without the use of gating system. When the residual motion was 3 cm, the results showed significant differences in dose distributions between the stationary condition and the moving phantom without gating beam control. The overdosed or underdosed areas enclosed about 33% of the treatment area. In contrast, the dose distribution on the moving phantom with gating window set to 0.5 cm showed no significant differences from the stationary phantom. With the appropriate setting of the gating window, the deviation of dose from the respiratory motion can be minimized. It appeals that limiting the residual motion to less than 0.5 cm is critical for the treatments of mobile structures.