The bidirectional reflectance factors vary as the incidence directions and the view angles change. At present the remote sensing is almost at nadir, therefore it is possible to improve the accuracy of remote sensing application by reasonably selecting the looking angle, solar zenith angle, and so on. Based on the multidirectional spectra of winter wheat canopy at several critical growth stages, the paper quantitatively analyzed the sensitivity of narrowband bidirectional reflectance to view planes, view zenith angle, solar zenith angle, growth stage, and band by using anisotropy factor (ANIF) and anisotropy index (ANIX). The change of NDVI with view zenith angle, solar zenith angle and growth stage was also studied. The results show that the anisotropy characteristics of bidirectional reflectance factors at solar principal plane was stronger than that at the other planes, and orthogonal principal plane was the weakest. The ANIX at solar principal plane was the biggest. The reason was that the shadow of canopy changed more dramatically at solar principal plane than at the other planes. The sensitivity of bidirectional reflectance factor at visible bands to zenith angles was stronger than in near infrared regions, the reason for which was that the shadow effect in visible regions was stronger than in near infrared regions. The ANIX in visible regions was bigger than in near infrared regions. The sensitivity of bidirectional reflectance factor to solar zenith angles increased as the view zenith angle increased. The NDVIs at every looking zenith angle all increased with the leaf area index increasing. The NDVIs at forward direction were larger than at backward direction, which resulted from that the shadow effect in visible regions was stronger than in near infrared regions. The solar principal plane implies rich internal structure information on object. In order to reduce the uncertainty from the observing method, the near infrared bands and small solar zenith angle should be chosen. The retrieve of structure parameters ought to select solar principal plane, and avoid hot spot region when inversing biological parameters using NDVI.