The invention provides a method for generating a random structure of a continuous fiber composite material and predicting the elastic performance of the continuous fiber composite material. The method comprises the following steps: generating a fiber model by a particle swarm algorithm, and optimizing the fiber model according to fiber jumping treatment to obtain particle space random distribution information for generating a three-phase RVE finite element model; and performing finite element simulation operation on the three-phase RVE finite element model to obtain a prediction result. Based on a representative volume element generation strategy of the particle swarm algorithm, the distance among fibers is controlled by the particle swarm algorithm in the process of generating the representative volume element, the requirement for the material volume percentage is fulfilled while fiber random distribution is ensured; and based on efficient generation of the representative volume unit and a homogenization theory, an elasticity prediction finite element model is established, a periodic boundary condition is applied, and the elasticity performance prediction result of the material can be obtained by microscomic finite element simulation, so that the efficiency of elasticity performance prediction is improved.