In healthcare systems, blood banks play a crucial role as central hubs for collecting, storing, and distributing blood and its components to meet the transfusion needs of patients. Blood banks face the critical challenge of balancing blood supply and demand to meet the needs of patients. This study presents a web-based intelligent system designed to enhance the efficiency of blood bank operations in Zakho District by providing a centralized platform for blood donors and blood bank staff to interact. The system aims to streamline blood donor registration, processing blood requests, inventory management, and forecasting future demand. The proposed system allows blood banks to add new donors, storage of their information and virus test results, and manage blood bag inventory to achieve this goal. The system allows blood bank administrators to send Short Message Service (SMS) notifications to a selected group of blood donors based on their blood group. The study evaluates two Deep Learning methods, namely Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU), using a dataset consisting of 2760 records of daily Red Blood Cells (RBC) demand from Zakho Blood Bank. Based on historical data, this study incorporates a Long-Short Term Memory (LSTM) model to predict future blood demand, including Red Blood Cells (RBC). The system provides a user-friendly interface for blood bank staff to manage the inventory and access donor information. It also includes features to post news and events for public engagement. The comparative evaluation of the LSTM and GRU methods revealed that the LSTM approach outperformed GRU. The LSTM model exhibited superior performance, yielding a Mean Squared Error (MSE) of 0.00217, Root Mean Squared Error (RMSE) of 0.046, R-squared (R2) of 0.918, and Mean Absolute Error (MAE) of 0.020. Furthermore, the system's usability was evaluated using the System Usability Scale (SUS), yielding a score of 82.3%, indicating a satisfactory level of usability.