Abstarct: Evolutionary algorithms are a subset of evolutionary computations and include the algorithms the search starts with multiple points in solution space to discover the answer and they are used in problems that ordinary solutions are not responsive due to complexity and various parameters of the problem. Evolutionary algorithms inspired by nature and the laws governing it try to calculate the optimal solution to difficult problems accurately. So far, various evolutionary algorithms are presented baxsed on the social behavior of organisms, biological behavior and physical phenomena. Pollination of flowers is an interesting process used by many plant species for reproduction and survival, and researchers have used the behavior of the plants to create an evolutionary algorithm called flower pollination algorithm. Finding the optimal solutions to an optimization problem is challenging and difficult by increasing the size of the problem by evolutionary algorithms such as flower pollination algorithm and the flower pollination algorithm may converge to local optimums instead of a global optimization which is solved by increasing the population and repetition of flower pollination algorithm which increases the running time significantly. In this study we try to use the members of the initial population parallel in the graphics processing unit which has higher processor cores than the main processor. The results of the tests and simulations indicate that the parallel version of the flower pollination algorithm by the parallelization technology of graphics (CUDA) has higher execution speed than standard flower pollination algorithm.