Abstract: Welding distortion is one of the critical issues that affect weldment quality and productivity.The existence of welding distortion severely affects not only the manufacturing process but also the service performance of welded structure.There are various kinds of welding distortions such as transverse shrinkage,longitudinal shrinkage,bending,angular distortion and buckling distortion.It is generally agreed that the welding distortion arises from the thermal ununiformity,thermal gradient in the thickness direction,localized incompatible plastic strain during welding and effect of residual stress after welding.The methods of welding distortion prediction are largely based on finite element analysis.In this paper,the newest research developments on prediction of welding distortion at home or abroad are summarized.Traditional finite element methods such as the thermal-elastic-plastic finite element method and the inherent stain method are introduced.Furthermore,the optimization design and artificial neural net (ANN) for welding distortion prediction that based on the traditional FEM are also presented.