Abstract: By formulating a more reasonable Error—In—Variables (EIV) model, total least squares method has been wildly used in recent years. Based on ill-posed least squares theory, coefficient matrixes are always ill-conditioned, and which will cause the results oscillate. The regularization is a well-known tool for solving these problems. In fact, total least squares method is a deregularizing procedure, so the ill-posed problems will be more serious. That means errors in the data are more likely to affect the total least squares solution than the least squares solution. In this paper, we propose using generalized regularization to solve ill-posed problems in total least squares, so as to improve stability of the results. Finally, numerical experiments are carried out to demonstrate the performance and efficiency of the generalized regularization method which have significant advantages in solving ill-posed problems.