The penetration of large-scale wind power has made the unit commitment of actual large power grid much more difficult to solve.Based on the scenario method，a dynamic reduction based multi-cut method is proposed.First，the stochastic unit commitment with wind farm integration is decomposed into the master problem and sub-problems.The master problem is used to determine the on/off status of generation units and the output scheduling corresponding to the forecasted wind power scenario，while the sub-problems are used to determine the generation scheduling corresponding to the sampling scenarios.Then，the master problem and the sub-problems are associated by optimal cuts and solved alternately.During the iteration process，the redundant optimal cuts with weak constraints on the master problem are dynamically reduced to make the number of optimal cuts remain constant in any iteration after its entry in a certain stage.Thus，the solution size of the master problem is reduced and the problem that the solution size of the master problem is abruptly increased with an increase of the iteration number is solved.Finally，a real provincial power grid including a large-scale wind farm and180thermal power units is adopted for calculation，the calculation results showing that the proposed multi-cut method with dynamic reduction is superior in terms of convergence and computational efficiency.