Abstract: In the process of grey clustering, the weights of indexes are always unknown and hard to obtain, and the decision paradox of the rule of "maximum value" often occurs. Aiming at the problems above, firstly, with the help of the Particle Swarm Optimization Algorithm, PSO-grey clustering coefficient vector is proposed to overcome the limitation of weights. Secondly, based on the theory of "entropy increasing theorem" and using the clustering weight vector group as an important tool, a multi-stage grey intelligent clustering model is established by introducing the entropy of clustering coefficient vector, which solves the decision paradox of rule of "maximum value" to a certain extent. To simplify the calculation process, the Matlab source code for this model is attached. Finally, by taking the drought risk assessment of irrigated agricultural areas in Henan Province as an example, the rationality and validity of the model are illustrated.