This project presents a complete Bayesian framework combined with the Random vector function-link nets (RVFL) algorithm for complicated data modeling, where we add the prior distribution both on the combination weights and the parameters of the basic functions.