Based on the least squares principle, we show how to estimate several different fuzzy regression models. Especially at the centre of our attention are the orthogonalisation of fuzzy regressors and models with orthogonalised fuzzy regressors. It is an extension in the fuzzy environment of the commonly known principal component analysis regression method. Rosset and Donzé [9] describe in detail the particularity of the methods. In complement to these first results, we investigate an orthogonal fuzzy least squares method based on Zadeh’s extension principle to solve fuzzy linear regression problems. Our method preserves the properties of the orthogonality, which is highly desirable. The effectiveness of the methods is shown in an empirical application. Fuzzy regression models are postulated and estimated with data from the SHARE Project (Survey of Health, Ageing and Retirement in Europe) [5].