Although the use of small segments in step and shoot IMRT provides better dose distribution, extremely small segments decrease treatment accuracy. The purpose of this study was to determine the optimum minimum segment size (MSS) in two-step optimization in prostate step and shoot IMRT with regard to both planning quality and dosimetric accuracy. The XIO treatment planning system and Oncor Impression Plus were used. Results showed that the difference in homogeneity index (HI), defined as the ratio of maximum to minimum doses for planning target volume, between the MSS 1.0 cm and 1.5 cm plans, and 2.0 cm plans, was 0.1%, and 9.6%, respectively. With regard to V107 of PTV, the volume receiving 107% of the prescribed dose of the PTV, the difference between MSS 1.0 cm and 1.5 cm was 2%. However, the value of the MSS 2.0 cm or greater plans was more than 2.5-fold that of the MSS 1.0 cm plan. With regard to maximum rectal dose, a significant difference was seen between the MSS 1.5 cm and 2.0 cm plans, whereas no significant difference was seen between the MSS 1.0 cm and 1.5 cm plans. Composite plan verification revealed a greater than 5% dose difference between planned and measured dose in many regions with the MSS 1.0 cm plan, but in only limited regions in the MSS 1.5 cm plan. Our data suggest that the MSS should be determined with regard to both planning quality and dosimetric accuracy.