The issue of households' residential choices is one of the key problem areas in urban economics today. While the first monocentric models focused on households' tradeoffs between the costs involved in location and the costs of travel between home and workplace, which was assumed to be in the economic centre of the city, more recent studies attempt to look beyond this simplifying framework and take into account current urban changes and the differences observed between European and North American towns. In particular, taking account of the role of urban and natural amenities improves the explanatory power of these theoretical urban models. Widely used in the field of environmental evaluation, the hedonic price method can provide useful responses to this urban problem. Developed by Rosen (1974), it uses household behaviours to measure the value that households place on different intrinsic and extrinsic housing characteristics, particularly natural or built amenities. Estimation of these hedonic prices must take account of the dual spatial and endogenous nature of the characteristics of housing: simple estimates by the ordinary least squares method are potentially biased. To correct this bias, the FGS2SLS method proposed by Fingleton and Le Gallo (2008) must be used, combining instrumental variables and spatial autocorrelation correction. This method is applied to the case of the town of Angers: the town centre has many historic amenities, and it is indeed a typical European example. Beyond the predominant role given to the intrinsic features of the housing, the way households value "green" amenities such as green spaces emphasises the complexity of the Angers urban space.