Objective To evaluate the feasibility and significance of CT image-based threedimensional (3D) brachytherapy for cervical cancer.Methods Three-dimensional (3D) plan and twodimensional (2D) plan were designed for 55 CT images of brachytherapy from 12 cervical cancer patients who received radical radiotherapy in 2013.Dosimetric comparison was performed between the 3D plan and 2D plan,and paired t-test,Wilcoxon signed rank test,Pearson correlation analysis,and Spearman correlation analysis were performed.Results A point dose,D90,V100,CI,and CI' in 3D plan were higher than those in 2D plan (P=0.015,0.016,0.000,0.000,0.000).Bladder point dose,rectal point dose,and rectal D2 cm3 in 3D plan were slightly higher than those in 2D plan,but hot spot dose was significantly reduced in 3D plan (P =0.140,0.123,0.214).Bladder D2cm3 was significantly higher than bladder point dose (P =0.000).Sigmoid colon D2cm3 was more correlated with the average doses of the three highest rectal points than rectal D2 cm3 (r =0.314,0.630,P =0.000,0.000).V100 showed a linear relationship with high-risk CTV (r =0.981,P =0.000).Bladder D2cm3 was higher than 430 cGy when the bladder volume was more than 80 cm3 ;small intestinal D2 cm3 did not change significantly when the bladder volume was less than 115 cm3,but decreased significantly once the volume exceeded the value.Conclusions Compared with the traditional 2D plan,the 3D plan for CT image-based cervical cancer brachytherapy significantly increases the target coverage and conformity index,but does not significantly increase the doses to organs at risk.Point dose evaluation is confirmed to be inaccurate.The doses to the bladder,rectum,and small intestine can be adjusted by controlling the bladder volume.