Purpose: The aim of this study is to evaluate the accuracy the collapsed cone convolution (CCC) algorithm for dose calculation in a radiation treatment planning system (TPS). Methods: We modeled various photon beams for various setup conditions in a radiation treatment planning system (CorePLANTM, Seoul C&J, Korea). The beam models were generated at various set‐up conditions such as open beam or wedged beam, 6 MV or 15 MV beam and field sizes from 4×4 cm2 to 40 × 40 cm2. Each beam model was optimized by spectrum modeling from measured percent depth dose (PDD) data, dose profile modeling from a measured profile at a specific depth (10 cm) data. Dose calculation was performed using conventional CCC algorithm. All measured data were acquired from a Clinac 21EX (Varian Medical System, Palo Alto, CA, USA) linear accelerator with the setting of SSD = 100 cm. All calculated PDD and dose profiles at various depths from generated beam models were compared to the measured data. Results: Calculated dose data from each beam model showed good agreements within 2% of difference to the measured PDD and within 3% dose profiles at various depths. Some regions such as penumbra region at 20 × 20 cm2 field size and horn region at wedge field showed dose discrepancies over 3%. The results of PDD at all situations showed well agreement with measured data under the 10×10 cm2 field size. For wedged cases, however, under the 5 cm depths, some inconsistency at penumbra region were appeared. Conclusions: In this study, we verified the accuracy of CCC algorithm in the TPS. Calculated results by our implemented algorithm was well satisfied with measured dose at small field size (〈20 7 times; 20 cm2). Our next study will perform to compensate theses inconsistencies.