In order to rapidly detect the chlorophyll content of winter wheat canopy leaves in the field, a vehicle-mounted multi-spectral imaging system with 2-CCD camera was developed, and the working performance of the system was analyzed at different vehicle speeds. The FOTON-4040 tractor was used as the vehicle platform equipped multi-spectral image intelligent sensing system. Four speeds were set up in field experiments, 〖JP3〗which were S1 (0.54 m/s), S2 (0.83 m/s), S3 (1.04 m/s) and S4 (1.72 m/s). Visible and near infrared canopy images of winter wheat were collected. Meanwhile, the GPS position information was obtained and the SPAD values which indicated the chlorophyll content of winter wheat leaves were measured. Multi-spectral images were processed by adaptive smoothing filtering and canopy segmentation. There were 10 parameters in the image detection. The average gray values of four bands ( R, G, B and NIR) were extracted, and four vegetation indices (NDVI, NDGI, RVI and DVI), mean value of H in HSI model and canopy cover degree C were calculated. The correlation between each parameter of the image and the SPAD value of the chlorophyll index was analyzed. The results showed that the correlations between the parameters of each image and the chlorophyll index at speed of S1, S2 and S3 were higher than that at speed of S4. The correlation coefficients between NDVI, RVI, NDGI and the SPAD value reached over 0.50 at speed of S1, S2 and S3. MLR models for the diagnosis of the chlorophyll content were established at different speeds of S1, S2 and S3, respectively. The model precision met the requirements of crop growing space distribution map. In order to further improve the diagnostic efficiency of the crops growth parameters in the field, the MLR model of the chlorophyll content in winter wheat leaves was built by NDVI, NDGI and RVI. The results showed that the model was universal. The research can provide support for the rapid diagnosis of field crop growth.