Three fields were captured with 6 MV photon beam, using a-Si EPID(SDD 100cm).Then the derived dose images were split into four regions and every region could be modified by using different methods to build models. The three dose comparison methods most commonly used in radiotherapy were used to analyze the absorbed doses in four regions. The result shows that the percentage dose difference comparison method is sensitive in steep dose gradient regions where small spatial offsets between two distributions may present large dose difference, while it works well in regions of relatively shallow dose gradients. In contrary, the DTA (distance-to-agreement) method is sensitive in shallow dose gradient regions where a large DTA value may be calculated for small dose differences, while it performs well in steep dose gradient regions. However, as the gamma method has referenced the merits of percentage dose difference and DTA, it performs well both in steep and shallow dose gradients and produces the lowest error. So the gamma comparison method can be performed better in clinical dosimetric verification.Cited