This paper deals with the problem of varying capacity load factors in master production scheduling. A simulation system of a production planning and control system and of a production system is described and an approach to compute the capacity loads from the simulation results is developed. This approach is used to measure period-specific capacity load factors for master production scheduling. The measured capacity load factors vary over time and show that the common assumption of the capacity load factors being period-independent is inappropriate. A stochastic programming model to cope with this variation is described. This model is compared to a common linear programming model for master production scheduling in a simulation experiment. The results show that the use of the measured capacity load factor scenarios improves the quality of the estimation of the capacity load factors significantly.