During the last years the multiple-model approach has become very popular and widely applied for estimation and control of stochastic systems under different uncertainties - unknown model structure or parameters. In the engineering applications different multiple model algorithms for system control have been proposed. The greatest number of them are of Bayesian nature. Their common feature is the bank of estimators providing separate state estimates required for the overall control synthesis. In the paper an Interacting Multiple Model (IMM) algorithm for stochastic systems control in the presence of parametric model uncertainty is designed. It is based on the cost-effective IMM estimator. The overall system control is synthesized as a probabilistically weighted sum of the control processes from separate regulators working in parallel. These regulators are synthesised for each model from the uncertainty domain. The regulators are based on linear system, quadratic cost function and Gaussian noise assumptions. The overall control process is computed as a state feedback. The cost effective IMM filter is used for partial state estimates generation. The algorithm presented is compared to other MM Bayesian algorithm for control through Monte Carlo simulation experiments. The simulation results demonstrate that the IMM control algorithm provides better results in the presence of abrupt changes in the parameters than the MMAC algorithm. The performance of both algorithms is comparable in a stationary mode.