The echo signals of airborne LiDAR based on streak-tube by neural network were classified. The typical echo signals of four different targets were compared and analyzed, and the intensity and morphology in echo signal were extracted as features. A BP neural network classifier were constructed by MATLAB. The effect of the number of hidden layer neurons, neural network training algorithms and training sample number on the performances of the classifier was compared and selected. The test results of echo signals using this classifier demonstrate that the accuracy of this classifier can reach 96% and Kappa coefficient is 0.95, which is capable to classify the echo signals accurately.