Because of its efficient identification performance, Iriscode has been widely used in iris recognition systems. With the emphasis on biosafety security, how to reconstruct the iris image from Iriscode to judge the safety of iris recognition system has become a research hotspot. In this paper, we analyze the problem of reconstructing iris images from Iriscode, construct an iris reconstruction optimization model and propose a two-stage fast iris reconstruction algorithm. We regard the process of extracting Iriscode as the dimensionality reduction process. In the first stage, some inexact iris reconstruction images are reconstructed according to the neighborhood of the given Iriscode using locally linear embedding algorithm. In the second stage, we use the particle swarm optimization algorithm and the inexact iris images obtained in the previous stages as the initial particle to perform the iterative search to search the accurate result. Experiments show that the reconstructed iris images obtained by the proposed algorithm can achieve the purpose of forging attack through the iris recognition system. Compared to other reconstruction algorithms, there is an increase in time consuming and accept rate of the reconstructed image.