Objective: This study investigated the dosage of computed tomography (CT) ionizing radiation at institutions affiliated to this medical group, compared the technical performance of different manufacturers, discussed the performance of different device models, and calculated the ionizing radiation dose as precisely as possible without affecting clinical practices. Methods: Of a patient population of 1,522 retrospectively analyzed for CT radiation dosage, 660 (43%) were female and 862 (57%) were male. We assessed three different CT modalities during the study period for the fulfillment of the eligibility criteria and recorded the CT dose index (CTDI) and calculated the size-specific dose estimate (SSDE) for each patient. Both imaging procedures and parameters have followed the recommendations and guidelines of the CT instrument. Regression analysis was applied to estimate the correlation between CTDI and SSDE. The index of offset dosage was determined with a regression equation. Results: The R^2 and slopes of the regression model between CTDI and SSDE showed that the automatic exposure control (AEC) technique substantially fitted the participant's body size in CT examination and reduced the dosage of radiation exposure (i.e., R^2= 0.79-0.96 and slope = 0.71-0.98 for thorax; R^2= 0.5-0.94 and slope = 0.44-0.97 for abdomen-pelvic). Meanwhile, the results of regression analysis showed that the 640-row CT was associated with minimum offset dosages. Conclusion: Our results demonstrated that the radiation dosage generated by 640-row CT was lower than that by the others, while the dosage of the 128-row equipment was higher than that of the 16-row. Therefore, the correlation between radiation exposure dosage and numbers of CT detectors was insignificant. Moreover, the minimum offset dosage (i.e., minimal difference between CTDI and SSDE) was associated with the 640-row CT. Hence, our analysis demonstrated that the 640-row CT could provide the most accurate radiation dosage for thoracic and abdomen-pelvic examination according to the patient's anthropometric parameters.