The conventional calculation model of spinning reserve is no longer applicable to the regional power grid group which contains small hydro and wind power access. Aiming at lower loss of load, higher utility of clean source and less cost of operation, based on the function models of the lowest thermal power fuel cost, the minimum expected outage expense, the least thermal power output fluctuation and the least abandoned water of main hydropower plants, a multi-objective optimization model is built considering the wind-water-fire power coordinated operation. The improved particle swarm optimization algorithm introducing particle concentration cognitive is used. By simulating and analyzing, the applicability and validity is verified. To compare them under different strategies, this method can get less fuel expenses of thermal power units in the case of lower loss of load probability (LOLP); optimize the spinning reserve capacity with the small hydropower and wind power output size. The model and the algorithm are helpful for drawing up optimization strategies of the reserve capacity in those regions of wind-hydro-thermal grid in which larger scale small hydropower and wind power exists.