A prototype of a biomass Stirling engine was developed and manufactured by an industrial company prior to the numerical investigations described in this paper. Dimensions and performance of the prototype had originally been estimated using the company’s own simplified “first-order” mathematical model of the engine’s working cycle. The manufactured engine was experimentally tested, and the results demonstrated that the power output from the machine was far less than expected. To understand what caused the engine’s low operational characteristics and to predict how to refine the design, more advanced numerical investigations of the working process of the engine were performed. This utilized a “second-order” type 5 control volumes model, together with three-dimensional computational fluid dynamics modeling. As an outcome of this study, several recommendations on to how alter the prototype’s design were forthcoming, which, in practice, allow significant improvements in the engine’s performance.