A mathematical model for short-term wind-hydro-thermal load dispatch is built up based on chance-constrained programming considering the stochastic characteristics of load and wind power. In order to make a better balance between local search and global search abilities, an evolutionary predator-prey strategy (EPPS) is used to solve the short-term hydro-thermal generation scheduling and short-term wind-hydro-thermal load dispatch problems. The simulation results show that EPPS can find better solution compared with other algorithms in short-term hydro-thermal generation scheduling. Meanwhile, the approach of chance-constrained programming provides the basis for setting reasonable spinning reserve capacity of system.