The accuracy of on-orbit absolute radiometric calibration of optical remote sensors directly affects the breadth and depth of quantitative applications, and thus the on-orbit vicarious calibration of the reflectance-based, irradiance-based, and radiance-based methods based on large and uniform sites plays an important role. However, due to a limited number of sites, low calibration frequency, low site reflectance, and single-point calibration cannot achieve full dynamic range calibration, the calibration accuracy is limited to 5%-8%. The improvement of the spatial resolution of optical remote sensors makes it possible for the targets-based absolute radiometric calibration with good spectral flatness and Lambert property. In this paper, the calibration principle, process and influencing factors of the gray-scale targets method are studied, and a simplified method for radiative transfer calculation is proposed. Furthermore, considering the response nonlinearity and dark current of high-resolution multispectral cameras, we apply the first-order function response model with bias to test a multispectral camera three time and obtain the calibration gain and bias, with the calibration uncertainty being better than 5%. In addition, the reflectance inversion validation is carried out on the laid color targets, and the results show that the absolute difference within 5%-70% reflectance is less than 0.01. In conclusion, the method proposed in this paper can realize the absolute radiometric calibration of optical satellite remote sensors in a large dynamic range and solve the common problem that the accuracy of radiometric calibration is generally not high in low-end quantitative applications.