One of the main issues of the single point incremental forming (SPIF) process is still the achievable accuracy. A number of methods have been suggested to increase this accuracy, but many of these contain a significant drawback. Reprocessing the workpiece can increase the accuracy but also significantly increases the manufacturing time and leads to a worse surface finish of the part. Other methods iteratively correct the toolpath based upon the deviations measured on the previously manufactured parts. This method is not very well suited for one of a kind products, since instead of one part, multiple parts need to be manufactured before the desired accuracy can be reached. Our method proposes to use feature detection to split the workpiece in a configuration of planes, edges, freeform surfaces and other features. For each of these features an optimised toolpath strategy can be determined and the toolpath in that zone can be adjusted for this strategy. The proposed method generates a single pass toolpath that leads to more accurate parts compared to the standard CAM toolpaths. This paper describes the feature based optimised toolpath generation method (FSPIF) and contains the results of experiments performed to validate this method.