In the optimization of unit commitment, it can provide a safer and more economical system operation scheme considering short-circuit current constraints and branch switching. However, its solution faces challenges due to the large scale of the problems. A heuristic decomposition algorithm is proposed based on analytical target cascade and large neighborhood search. First of all, the joint optimization model of unit commitment and branch switching incorporating short-circuit current constraints is established and hence it is transformed into a mixed integer linear programming model with a separable structure. Then, the analytical target cascade method is used to decompose the converted model into upper level coordination master problem, lower level integer programming and linear programming subproblems. Moreover, the large neighborhood search is introduced in solving the lower level subproblems. Finally, the simulations on IEEE 118-bus and 54-unit systems and an actual power system show that the proposed algorithm can quickly converge to feasible solutions, whereas not affecting solution quality.