[en] Objective: To assess PET reconstructed image effected by different ¹⁸F-FDG dosages with quantitative and qualitative analysis. Methods: To perform PET phantom acquisition by routine clinical parameters after filled with different doses of ¹⁸F-FDG solution. An identical slice was extracted from reconstructed image for doing following analysis: the hot area standard uptake value (SUV), the ratio of hot area to cold area, the standard deviation on background area, the ratio of true coincidence to random. Results: 296 MBq: The image uniformity was terribly worse, T/R=0.83, other indexes were irregular. 148 MBq: The image presentation looked like the image without attenuation correction, T/R=1.64, other indexes were moderate. 74, 37 and 18.5 MBq: The images were with excellent uniformity, resolution and contrast, the background noise was suitable, all of the quantitative indexes were good. 9.25 and 4.625 MBq: The uniformity and resolution was degraded terribly because of the higher noise and lower information. Conclusion: Combining above results with other considerations, such as radiation exposure, information amount and acquisition time, the authors think the optimal dosage should be 4.625-11.1 MBq/kg