The supply chain performance depends on accurate demand forecasting. This becomes more critical when it comes to non-contract spare parts service supply chains. This is because of the fact that customers are not obliged to place an order for the required spare parts to its Original Equipment Manufacturer (OEM) due to the availability of multiple suppliers. The business aircraft spare parts supply chains are the ones most affected by this phenomenon because their travel pattern and usage is totally unpredictable in comparison with passenger airline carriers. These highly uncertain and unpredictable demands and subsequent inaccurate forecasts have severe financial consequences. It is also computationally expensive to predict demand forecast for each part due to huge number of spare parts in business aircraft's supply chain. Hence, in this paper the objective is to investigate forecasting methods, their variants and artificial intelligence (AI) methods, developed for irregular demands, to propose best method variant that is capable of accurately forecasting not only uncertain but unpredictable demand e.g. business aircraft's spare parts supply chain. We retained Boot Strapping (BS) method as the most suitable base method for uncertain and unpredictable demand forecasting. This is because of its inherent ability to reduce error due to resampling with replacement. The point and interval (existing), and sliding window (proposed) BS methods are implemented in Matlab and results of demand forecasts are compared with the forecasts generated from benchmarked existing forecasting methods as: Croston, Croston variants (SBJ, SNB, TSB), moving average (MA), single exponential smoothening (SES) and Commercial (proprietary black box) methods. The data used in this study is collected from Dassault Aviation. The results demonstrate that proposed sliding windows BS variant with ‘Mean’ function outperformed 75% of the spare parts with significant financial gains in terms of inventory holding and shortage costs. © 2017