In practices, measuring leaf area index (LAI) in large area scale is very difficult. Therefore, retrieving LAI quantitatively based on remote sensing technology is concerned by many researchers. We proposed a BP-ANN based method to retrieve winter wheat LAI using surface reflectance data of MODIS. The MODIS pixel is assumed to be composed by crop canopy and bare soil. The SAILH (Light Scattering by Arbitrarily Inclined Leaves including the Hotspot-effect) model was used to simulate the directional reflectance of crop canopy, and the bare soil was assumed to be Lambertian. Series LAI maps of winter wheat in Shunyi District, Beijing were retrieved using this method during April in 2001. The research indicated that this method can be used well to retrieve LAI in large area scale, which is valuable to monitor crop growth.