This paper established the two-stage operation optimization model, aiming at the power system day-ahead operation problem with wind-storage system in power market environment. In the first stage, based on the probabilistic scene set which describes the uncertainty of wind power, considering the wind curtailment and bidding deviation, taking the maximum expected income of the wind-storage unit as the goal, the day-ahead bid quantity of the unit is optimized. In the second stage, considering the factors such as wind-storage unit force deviation penalty, environmental cost and network constraint, and aiming at the minimum comprehensive operation cost of the system, the day-ahead unit commitment is optimized. In addition, this paper proposed a deviation penalty mechanism for day-ahead clearing. The simulation experiment of IEEE-30 bus 6-machine system shows that compared with the traditional scheduling model, the model has the characteristics of less wind-storage dispatching deviation, lower comprehensive operation cost of the system, and obvious environmental protection benefits.