In order to improve the cost and quality advantages of feed enterprises, more scientific methods can be used to formulate production scheduling plans. Firstly, according to the characteristics of feed processing scheduling, a scheduling model based on batch organization production is constructed. Secondly, aiming at the problems of slow convergence speed and weak local search ability of cuckoo algorithm, an improved cuckoo optimization algorithm with different improvement strategies is formed. The NEH method, the Logistic chaotic mapping method and the random method were employed to create initial solution. The strategy of dynamically changing the step size is used to balance the exploration ability and exploitation ability of the algorithm. The crossover stage based on differential evolution is added to enhance the mining ability of the optimal solution. The improved cuckoo search algorithm is used to solve 40 Taillard test problems and an actual feed scheduling problem with minimization of total flow time. The experimental results show the effectiveness of the proposed algorithm in solving flow shop scheduling problem.