The reliable prediction of process-induced fiber microstructure, especially for long fiber reinforced thermoplastics (LFRT), is still a considerable challenge. The prediction by means of computational models with standard settings is not effective and current simulation environments do not offer the possibility to optimize the resulting fiber orientation quickly. In case of distinctive anisotropy, the consideration of fiber orientation within the scope of structural mechanics in part design can be considered as indispensable. The basis for an improved prediction is a new developed optimization approach for an automated calibration of the resulting fiber orientation tensor by means of certain experimental tests. This approach allows to optimize the associated model parameters within a few minutes by calculating a single initial flow field.