Abstract: The morphological feature of plasma jet was obtained by digital image processing and pattern recognition,and characterized by the aspect ratio.To develop a knowledge base for the intelligent control of plasma spraying process,the cluste-ring analysis was carried out between spraying parameters and the feature of plasma jet based on data mining technology.The quali-ty of clusters was evaluated by the properties (residual stress,porosity and hardness) and microstructure of the coatings,and the inference rule sets from processing parameters to the high quality cluster were created.Hence,the investigation of the rela-tionship between the processing parameters,plasma jet and coat-ing performance was realized.Finally,an intelligent control approach for the sub-system of plasma jet generation was presented in this paper based on the rule sets supervision and multi-model adaptive control system structure.