AbstractA new technique for nonlinear state and parameter estimation of discrete time stochastic volatility models is developed. Algorithms of Gibbs sampler and simulation filters are used to construct a simulation tool that reflects both inherent model variability and parameter uncertainty. The proposed chain converges to equilibrium enabling the estimation of unobserved volatilities and unknown model parameter distributions. The estimation algorithm is illustrated using numerical examples. Copyright © 2002 John Wiley & Sons, Ltd.