Abstarct: In the existing views of fuzzy regression analysis, the improvement of the model parameter estimators has been only from the perspective of correction of the existing meters and improving the problems of the distances and previous meters. While in classical (nonfuzzy) regression, the estimators are improved in order to achieve the optimality of the comparison criteria and the goodness of fit is also done by manipulating and generalizing the estimators themselves. Therefore, in this study, we aim to improve the fuzzy optimization criteria by improving the estimators in fuzzy regression models, and not the desired meter. In this regard, we propose the use of shrinkage estimators, resampling methods, and penalized methods in fuzzy regression models and show how fuzzy optimization criteria can be improved by using fuzzy improved estimators ( shrinkage and penalized methods) or retrieval of fuzzy data