In the era of big data, data is increasingly complex, often containing outliers, fuzzy data, etc. In this paper, a new weighted least square method is proposed to estimate the parameters of the fuzzy linear regression model whose input are exact numbers, output and regression coefficient are trapezoidal fuzzy numbers. By using simulated data sets and real data sets, and compared with other methods in fuzzy regression analysis, it shows that the estimation method proposed in this paper is more effective.