To address the challenge of accurately recognizing signals in real time for space debris laser ranging, this paper proposes a deep learning-based method for real-time recognition of space debris laser ranging signals. Utilizing the long short-term memory (LSTM) network of recurrent neural networks in deep learning, the proposed method enables real-time recognition of signals that are difficult for traditional methods to detect. The LSTM network excels at capturing and maintaining long-term dependencies in time series data, effectively handling missing, noisy, or irregular sequences, while also demonstrating strong generalization capabilities. Validation using actual observational data shows that the proposed method improves the recognition rate of space debris. Compared to the echo signal correlation method, the average running time is reduced by ~8%, and the F1 score for signal recognition increases from 0.2144 to 0.6068, representing an improvement of nearly twofold. This method will play a significant role in the detection and identification of space debris and provide valuable insights for the application of deep learning techniques in space target laser ranging.