Abstract: GF-1 satellite has the characteristics of high spatial resolution and short revisiting period, serving as the first satellite of the High-resolution Earth Observation System for National Science and Technology Major Project in China. The satellite carries two multi-spectral high-resolution cameras (panchromatic multispectral sensor, PMS) and four multi-spectral medium-resolution camera (wide field of view, WFV). GF-1 captured data play an important role in the identification of the underlying surface, and these data can be obtained free of charge from the website of the China Centre for Resources Satellite Data and Application. An important step for the application of GF-1 satellite data can be the interference removal of atmospheric molecules, aerosols, ozone, water vapor. The spectral response curves from Landsat8 (LC8) operational land imager(OLI) and GF-1 MSS/WFV were then analyzed in the visible and near the infrared bands. The results showed that the spectral range of LC8 OLI in the red- and near the infrared bands was relatively narrower than that of GF-1 MSS/WFV, whereas the spectral response function in the blue- and green bands was slightly different from that of GF-1 MSS/WFV, indicating that it is feasible to transplant LaSRC correction process to GF-1 MSS/WFV data. Since GF-1 satellite lacks the short-wave infrared band compared with LC8, the algorithm was modified to adapt to the characteristics of GF-1 channels. The atmospheric correction project was designed for the GF-1 satellite MSS/WFV data, including the algorithm analysis and code implement based on 6S atmospheric radiation simulation model and C++ programming language. Some parameters were used to estimate initial aerosol, including total atmospheric transmission, gaseous transmission, atmosphere spherical albedo and actual values of digital elevation model, atmospheric precipitation, ozone content. The loop calculation of the aerosol optical thickness(AOT) was carried out until the ratio between the red- and blue bands of GF-1 MSS/WFV data equal to the prescribed ratio of MODIS, according to the relationship between the blue- and the red surface reflectance known from MODIS. The results can be obtained the surface reflectance with the minimum residual error during different ?ngstr?m coefficients, and retrieved the aerosols in the pixel level of GF-1 MSS/WFV data. The pixel aerosol and these parameters were then substituted into 6S model to calculate the surface reflectance. Since the project was equipped a data file containing the atmospheric precipitation and ozone content at the current day, the surface reflectance could be obtained when only inputting GF-1 MSS/PMS data. Because the data of atmospheric influence gases, such as ozone and water vapor, on the same day of the data to be corrected were sometimes difficult to identify, two schemes can be provided, one is to use the data at that time, the other is to use the daily values of ozone and water vapor in past six years instead. The experimental results show that the proposed method has a good effect on the atmospheric correction in the middle and low latitudes that covered by vegetation, such as farmland and trees, but not good effect on that of the bare land and building surface that covered by sparse vegetation. Based on 6S model and LaSRC correction process, the correlation coefficient of the atmospheric correction between GF-1 MSS/WFV and LC8 OLI was from 0.825 to 0.972, indicating a high correlation of atmospheric correction results for two satellites. WFV similar spatial resolution to that of LC8 OLI was in good agreement with that of LC8 OLI atmospheric correction compared with that of MSS. The results show that it is convenient and operable for the GF-1 satellite data atmospheric correction method using the self-estimation aerosol parameters in the pixel level based on 6S model and LaSRC process. This promising atmospheric method can be very suitable for the land surface application, such as agricultural and forestry monitoring in growing season. At present, this method has been successfully implemented on Remote Sensing data processing platform Remote Sensing Desktop (RSD) in China.