A new optical instrument for fast determination of pear leaves nitrogen status was designed and fabricated. A multi-spectral imaging system was used as optical detector. In the paper, the principle of multi-spectral imaging for the measurement of leaves nitrogen status was first introduced. Then, the method of using error back propagation artificial neural network (BP-ANN) for calibration modeling was elaborated. Mean reflective light intensities in all images covering blue, green, red and infrared wavelength were taken as the input data to BP-ANN. The structure of BP-ANN with three layers has been optimized to minimize its calibration error. In the test, total 200 leave samples picked from Huang-hua pear trees planted in three orchards with different nitrogen fertilizing schemes. Among them, 150 samples were selected randomly out as for calibration set with the remaining 50 for prediction set. The result shows that correlation coefficient of R2 between predicted and measured values of nitrogen content reaches 0.82 with maximum prediction error less than 4.72(SPAD). The study suggests that the new optical method integrating multi-spectral images with BP-ANN is promising for fast diagnosis of fruit tree nutrition status.