The pultrusion process is one of the most efficient production technologies for fiber reinforced polymers. The continuous process has been established for decades, however, not all relevant process relationships are sufficiently understood yet. This work introduces a data-based and multidisciplinary process analysis approach by utilizing simulations as well as experiments with application of an inline quality assurance system as combined data source. The approach enables correlation analyzes of synchronized data sets from both domains, which leads to an extended information linkage compared to conventional methods. A transient process model combined with an epoxy material model is implemented in a finite element software and extended by closed-loop process control algorithms. The pultrusion line was equipped with sensors for die temperature and die-cavity pressure, pull speed and pull force, ultrasonic testing and surface quality. Two studies are carried out on a glass fiber reinforced profile. The first study investigates steady process states by experiments and simulations. Among others a criterion for the minimum degree of cure was identified that is required to obtain flawless profiles. The second study examines the influence of different die designs, control approaches and process parameters on the transient process startup by simulation. The results show that the die design and the temperature control strategy significantly affect the startup time until a steady-state is reached.