In the multi-response optimization design, the model parameter uncertainty and the noise factors in the production process inevitably lead to high fluctuation of the predicted response value. In view of the above problems, a method for multiresponse optimization is proposed based on the Bayesian sampling technique, the Pareto optimization strategy and the grey relational analysis method. Firstly, a Bayesian multiple regression model is used to construct the relationship between the process responses and the experimental factors when considering the uncertainty of the model parameters. Secondly, the Pareto optimality frontier is obtained based on the Pareto optimal strategy, and then the Bayesian posterior probability for experimental points on the Pareto optimality frontier is calculated. Thirdly, the grey relational analysis method is used to identify the optimal design scheme. Finally, a practical example reveals that the proposed method can improve the robustness and the reliability of optimization results when considering the variability of the predicted responses.