This study presents a time series forecasting approach for predicting student admissions to higher education institutions in Azerbaijan utilizing the Long Short-Term Memory (LSTM) model, a type of recurrent neural network (RNN) particularly effective for sequential data. As demand for higher education continues to grow, accurate forecasting is essential for informed policy-making and resource allocation. The dataset used spans from the academic year 2000/2001 to 2023/2024 and was preprocessed to handle missing values and ensure temporal consistency. The LSTM model was implemented in Python using TensorFlow and Keras libraries. Data were normalized using MinMaxScaler, and a one-step-ahead forecasting approach was adopted, where previous years’ admission counts were used to predict the subsequent year. The model was trained on historical data and used to forecast admissions for the academic year 2024/2025. Results show that the LSTM model is capable of capturing underlying temporal patterns and trends in student admission data. The forecasted admission count for 2024/2025 is 56,247 students, which aligns closely with recent years’ trends, indicating the model’s reliability. This approach illustrates the potential of deep learning methods like LSTM in the domain of education analytics and can support decision-makers in planning and policy formulation.