The multifractal characteristics of thyroid dynamic infrared images of healthy subjects were studied in this paper, and the statistical analysis and differential test of thyroid multifractal characteristics of different individuals were performed. Firstly, in the constant temperature and humidity experiment environment, multiple frames of human thyroid infrared images were acquired and meshed to form a temperature-time series. Then, the original signal length, wavelet transformed scale factor and statistical moment order of the multi-fractal analysis of human thyroid were discussed. After determining the above parameters, the multi-scale wavelet transform of the temperature-time series was performed to solve the Wavelet Transform Modulus Maximal, and then the distribution characteristics of multi-fractal characteristics of the left and right thyroid lobe of different healthy subjects were solved. The results show that the singularity index c1 of the thyroid multi-fractal characteristic line dimension of healthy subjects concentrates in the range of 1.1-1.3, and the gap coefficient c2 concentrates in the range of 0.002-0.005. The test level of individual differences is α=0.01; the half-width of the multi-fractal line concentrates in the range of 0.164-0.166 and the test level of no individual difference α is 0.05.