Cybersecurity is a field aimed at protecting digital systems and networks from malicious activities. With the growing complexity and sophistication of cyber threats, there is a need for advanced techniques to enhance the effectiveness of security measures. Machine learning models have emerged as a promising approach in cybersecurity, enabling the development of intelligent systems capable of detecting, preventing, and mitigating various cyber-attacks. This paper provides an implementation of deep learning model long short-term memory (LSTM) for classifying network attacks both for binary classification and multi-classification. The NSL-KDD dataset, a modified version of the KDD Cup 1999 dataset has been used for the experiment in binary and multi-classification tasks of network attack. LSTM models well performs at capturing and modeling temporal dynamics in data. They can learn and exploit patterns over time, which is essential for network attack classification since many attacks exhibit specific pat terns or behaviors in the network traffic. The performance metrics like recall, precision, F1-score are analyzed and a comparative analysis done with various classifiers. The model has been fine-tuned with proper hyperparameter settings and showed an improvement in performance metrics. There are research efforts to address various challenges and propose potential solutions.