Pinus species hold important ecological and economic value. However, the large genomes and slow molecular evolution of Pinus species result in highly similar morphological traits, complicating inter-specific identification. In order to solve the problems of high cost, long time, low accuracy and complex operation of traditional identification methods, this paper designed a detection method based on near-infrared spectral data (NIRS), one-dimensional continuous convolutional neural network (1D-CS-CNN) and Auto-encoder technology for Pinus species. Initially, the 1D-CS-CNN model uses the efficient Continuous Structure (CS) to replace the hidden layer structure in the traditional 1D-CNN model. The model can be directly applied to analyze one-dimensional near-infrared spectral data (NIRS). Next, combining the reconstruction error of the auto-encoder, a identification method considering an unknown origin is designed for Pinus species, which can solve the problem of high confidence in convolution neural networks by using an auto-encoder and reconstruction errors of the samples to be tested. Whether the sample is an unknown variety can be determined by comparing the corrected confidence level with the preset threshold value. The results show that the 1D-CS-CNN training set achieves 100% accuracy, with the loss value stabilizing at 0.015. Compared with the traditional 1D-CNN model, the improved 1D-CS-CNN model has faster recognition speed. Meanwhile, the accuracy rate of the auto-encoder for the category detection mechanism of Pinus species from an unknown origin is 99%. The experimental results show that the model can quickly and efficiently classify different species of Pinus and detect the new species in Pinus.