The traditional atmospheric radiative transfer calculation method has been unable to meet the needs of space-borne hyper-spectral infrared atmospheric remote sensing data processing because of the limitation of the computing resources and efficiency. Based on the optimal spectral sampling method, this paper developed a fast and accurate high spectral resolution infrared atmospheric radiative transfer model FFRTM_IR. This model was used to simulate the measurement of hyper-spectral infrared radiance atmosphere sounder (HIRAS) aboard on FY-3D satellite. Independent profiles validation results show that the bias of FFRTM_IR were less than 0.06 K and the standard deviation were no more than 0.1 K for all HIRAS channels. Under the same calculation environment, the speed of FFRTM_IR was slightly faster than the general radiative transfer model, such as CRTM. Along with the FFRTM_IR model, an analytical method was used to drive the temperature, water vapor, carbon dioxide and ozone profiles jacobian matrix, which agree well with results obtained from accurate perturbation method. All these validation and analysis results showed that the developed initial efficient infrared atmospheric radiative transfer model can be used in space-borne hyper-spectral infrared atmospheric sounding instrument simulation and data processing.