Throug h using modern sensor technology and computer monitoring technology,a real-time monitoring system of electrical and mechanical signals was established to monitor welding current,voltage between electrodes,dynamic resistance and dynamic electrode force during resistance spot welding.According to the regularity of dynamic resistance and electrode force,characteristic values were extracted from these signals by correlation analysis. A welding quality prediction model was established by using multiple linear regression analysis. Preset welding current,preset electrode force,welding time,charac-teristic values extracted from dynamic resistance and dynamic electrode force were chosen as the input variables for the model,and nugget diameter was chosen as the output variable.The results showed that the established model could forecast nugget diameter accurately and the maximum deviation was no more than 0.4 mm.