Abstract: The weld metal expulsion and sticking electrode are the main factors which cause the occurrence of substandard appearance quality of joints in resistance spot welding.The digital images obtained from the appearance of welding joints were used as sources of information.First,through the analysis of images in which expulsion and sticking occurred during the welding process,the perimeter L,area S₀,elongation A and density C were selected as the parameters to reflect the characteristic of the binary image of the joints.Second,the law between the four parameters and welding parameters was revealed based on a lot of experiments.The three parameters of L,S₀ and A were extracted as the characteristic parameters to identify appearance defects of the joints.On the basis,an evaluation model was established for the appearance defects of the joints based on support vector machine.At last,the actual verification results showed that the evaluation model can diagnose the appearance defects caused by expulsion and sticking electrode,and its accuracy can reach up to 96.67%.