Abstract: The principal component analysis was used to extract the features of indicator factors, the grey prediction method was applied to construct the prediction indexes, and the MATLAB software and BP neural network were used to predict the change of area of cultivated land in Xi[JP8]’a[JP]n. The results showed that the BP neural network was characterized by the simple structure, fast convergence and high precision after analyzed with the principal component analysis. The accuracy of predicted annual reduction of area of cultivated land with the BP neural network was high, and the BP neural network was reliable and feasible. The predicted results showed that the area of cultivated land in Xi[JP8]’a[JP]n in 2013 was 248 826.67 hm².The reduction of area of cultivated land in Xi[JP8]’a[JP]n was mainly caused by the rapid economic development and the increase of urbanization level. It was suggested to pay great attention to the planning and protection of cultivated land resources in Xi[’an in the urban development.

Key words: principal component analysis (PCA), BP neural network, prediction of cultivated land, Xi’an