Minimal attribute reduction (MAR) in rough set theory is a NP-hard nonlinear constrained combination optimization problem. In order to improve the solution quality and computing efficiency of rough set minimum attribute reduction algorithms based on swarm intelligence, this paper proposed a multi-strategy improved pelican optimization minimum attribute reduction algorithm (POAMAR). Firstly, the population gathering strategy was proposed to improve the initial state of the population individuals and make the individuals perform well. Then, opposition-based learning was introduced in the iterative process, and adaptive opposition-based learning was designed to improve the poor individuals in the population, expand the search space and reduce the amount of calculation. Finally, in order to avoid a large number of repeated calculations of the fitness function, a list search strategy was proposed to quickly calculate the fitness value. Experiments are carried out on six UCI data sets and compared with four representative algorithms. The results shows that the proposed algorithm performs well in reduction ability, running time and convergence speed, and has certain advantages in solving the minimum attribute reduction problem.