A stochastic programming model system is formulated for the control of a discrete time parameter Gaussian process. The problems representing the model system are solved by nonlinear programming procedures. The method can be applied e. g. for the dynamic regulation of the water level of the Lake Balaton, Hungary.