ESWAT is a simulator that integrates catchment and river water quantity and quality processes. The integration leads to a high number of model parameters, which complicates model calibration. As the model is semidistributed, the water quality and quantity variables at different observation sites inside the catchment can, and should, be used during this process, in order to use all the available information. A simultaneous use of all the different observed series and a high number of free parameters, however, creates a complex mathematical problem. Existing methods such as Pareto-optimization are practically very difficult, if not impossible, to implement. We present therefore a new methodology that reduces the many objective functions to a single global criterion in an objective way, excluding the weighting problem. The global criterion then is minimized using a global search algorithm, i.e., the shuffled complex evolution method. The methodology is applied on the Dender River basin (Belgium), a heavily modified river basin with irregular flows.