This paper puts forward an adaptive mutation of the quantum particle swarm optimization (AMQPSO) algorithm in order to improve the efficiency of autogenerating test paper. Firstly, a method of effective premature and stagnation judgement is embedded in the algorithm. Once premature signs are retrieved, the algorithm mutates particles to jump out of the local optimum particle according to the structure mutation. Secondly, the algorithm constructs a mathematical model of autogenerating test paper in steps based on Item Response Theory to reduce redundancy and improve the efficiency of autogenerating. Simulation results showed that compared with the genetic algorithm, the proposed algorithm is of better performance in both success rate and quality of autogenerating test paper.