Real time estimation for the quality of the resistance spot welding (RSW) is very important in the quality control of the automotive industry required. In this study, the real time estimation for the quality of RSW was tried using an artificial neural network. The input variables the artificial neural network were material information, welding voltage, welding current, and electrode force. The output variables were tensile shear strength, nugget size, and fracture shape. The material used were 780 MPa to 980 Mpa, and the material combination was made of the combination used in the automative body. The verification was performed and compared with the welding quality prediction result. As a result, it was confirmed that the prediction error of tensile shear strength, nugget size, and fracture shape was less than 5%. It is confirmed that the artificial neural network algorithm developed in this way enables the real time welding quality prediction.