Abstract: To overcome the shortcoming of the traditional PLS algorithm, a discounted-measurement recursive PLS (DRPLS) algorithm is proposed to build prediction model of the Mooney-viscosity of rubber mixture. This improved RPLS model could be on-line updated with new data, and more adjustable parameters are available. On the other hand, the rubber-mixing process is a typical non-linear batch process with very short operation time, and the rigorous mixing mechanism is not well developed. The Mooney-viscosity of the mixture is the key factor that has great effect on the quality of rubber products. By this developed DRPLS algorithm, a highly precise and flexible Mooney-viscosity model could be built. Application results confirm superior predictive precision of the developed approach.