Abstract. The sludge discharge model in wastewater biochemical treatment is highly nonlinear,Accurate prediction of sludge discharge will provide basic data for the precise control of subsequent sludge dewatering process。In the research based on this method, due to the relevance vector machine(RVM) has high sparseness and uses probability factor in predict, which is superior to the support vector machine(SVM) in the sludge discharge. However, the kernel function parameters of RVM are estimated by experience. Therefore, a kind of RVM method based on the particle swarm optimization(PSO) algorithm is proposed, which adopts the PSO algorithm to determine the kernel parameter of RVM, then builds RVM model and uses forecast the final sludge discharge.