With the technical development of wind power forecasting, making wind power generation schedule in power systems become an inevitable tendency. This paper proposes a new dispatch method for wind farm (WF)cluster by considering wind power forecasting errors. Aprobability distribution model of wind power forecasting errors and a mathematic expectation of the power shortage caused by forecasting errors are established. Then, the total mathematic expectation of power shortage from all WFs is minimized. Case study with respect to power dispatch in aWF cluster is conducted using forecasting and actual wind power data within 30 days from sites located at Gansu Province. Compared with the variable proportion method,the power shortage of the WF cluster caused by wind power forecasting errors is reduced. Along with theincrement of wind power integrated into power systems,the method positively influences future wind power operation.