Abstract: In agricultural diseases forecasting system, it usually uses the method of expert group decision making and the expert support degree to reflect the effect of expert personal decision to the group decision consequence. In this way, the rationality of the group decision consequence can be improved effectively. In this kind of systems, the static expert support degree is always employed, and the expert’s personal decision has randomness because of the different environment. Therefore, the randomness could lead to a wrong result. It presented a dynamic method of expert support degree, and set up the model of multi-expert group decision for the agricultural diseases forecasting system, and used the dynamic weight value calculation method to adjust the expert support degree in this system. The method can not only reflect the experience of expert, but also resolve the question from the randomness of the expert’s views, and improve the scientificity and the reliability of the group decision making of the agricultural diseases forecasting system. This method has been used in practical project.