<p>After the occurrence of severe meteorological disasters such as floods, typhoons and hurricanes, a large amount of water is often brought, which will directly hinder the process of disaster relief. Therefore, rapid drainage is essential to save the affected people and the means of production. Pump is the key equipment for waterlogging drainage, but the drainage capacity of a single pump is weak, and rapid drainage requires multiple pumps to work in parallel. However, the post-disaster drainage site often faces the challenge of material and energy shortage, but it is difficult to deploy a large number of pumps of the same type, and the power resources that can be used are very limited, so it is necessary to consider the feasibility of parallel use of different types of pumps and optimize the control strategy as far as possible to reduce energy consumption. In this paper, an artificial bee colony algorithm based on improved search equation is proposed, which has stronger local search ability, accelerates the convergence speed, improves the robustness of the algorithm and can optimize the control of different types of parallel pumps, so as to achieve the optimization of energy consumption. In order to verify the feasibility of the method, this paper constructs a mathematical model with the objective of minimizing energy consumption, uses the improved artificial bee colony algorithm to simulate and solve the control scheme under different working conditions and builds a test bench for experiment. The number of pumps obtained from the optimal control strategy under the test condition is consistent with that obtained from the simulation results of the control strategy, which confirms that the algorithm has good optimization performance. It is of great significance in the efficient use of different types of parallel pump sets and also verifies the feasibility of the weighted normalization method in the actual application of pump sets.</p>