The efficient operation of blood transfusion services is vital due to the fluctuation in the demand-supply of blood components, which is critical in saving patient lives on a day-to-day basis. Blood inventory faces issues of unpredictable demand, shortages and wastage which can be addressed by balanced collection and distribution helping create a robust blood supply chain. Healthcare institutions require realistic demand forecasts to assist in the development of a decision-support system based on donation data obtained from the blood bank. The research forecasts blood donations using time series and machine learning models based on synthetic univariate data that simulates real-time blood bank donations. The effectiveness of the models utilized is evaluated using regression metrics. Experiments are performed to better understand the nature of predictions based on trends and patterns in time series data. Further work can be based on the other information in the dataset, leveraging multivariate data for analysis that can assist improve the blood supply chain’s robustness in several ways.