The present paper discusses a new method of classifying kinds of material and driving stages and styles of Volvo wheel loaders (WLO). This is achieved by indirectly monitoring relevant, but usually latent variables, based on directly monitored sensors of WLOs. The continuous classifications will support Volvo's actual objectives such as, e.g., maximizing the remaining useful life of components, fuel efficiency, and productivity. To this end, a set of WLO machines was equipped with extra sensors and collected a limited dataset with richer information. Based on this limited dataset, different machine learning (ML) methods were tested to derive and to verify the classifications. It showed that support vector machines (SVM) produced the best results: the driving styles could be classified with a test accuracy of 77% (resp, 99.5%) in the loading (resp. unloading) driving stage. Further the SVM model is also verified both theoretically to enhance the confidence in the model and experimentally with a set of additional test drivers.