Rubber is the commodity of the results of demand levels andagricultural production always has increased significantly from time to time.This is due to the high demand of the company's suppliers are the result ofprocessed rubber to meet the needs of production. However, due to theinfluence of the conditions of the global economy so that it appears theinstability of prices. The data used in this research in the form of aUnivariate time series data is converted into the multivariate. The methodused is the method of Back propagation Neural Network (BPNN) is appliedto the data time series technical specific rubber commodity prices with thehelp of weighted optimization Particle Swarm Optimization (PSO) whichhopefully may help to improve the performance of the prediction so thatresults of the RMSE for the prediction of rubber prices gained can be moreaccurate. Of research results obtained the best model on a back propagationneural network with the parameters for the training cycle 600, the learningrate and momentum 0.1 0.2, as well as neuron size 3 whereas in particleswarm optimization value of population size 8, max value. of generation 100,the value of inertia weight 0.3, the value of the local best weight 1.0 andglobal best value weight of 1.0 produces a better RMSE value i.e. 0.040compared to just using the BPNN alone i.e. 0043. This proves that the PSOmethod able to give better results.