We propose a generic methodology to link landscapes and urban sprawl. The objective is to identify landscape features that can explain the urban sprawl. Our methodology first quantifies the landscapes and the urban sprawl using indices computed from multi-temporal land use maps. The indices are then statistically analysed using principal component analysis (PCA). PCA highlights the indices correlations and it especially identifies the landscape features that prevent or foster urban sprawl. We present the results of this methodology for the metropolitan area of Angers (France).