The purpose of the present study is twofold: the first objective is to evaluate the importance of visual experience for the ability to form a spatial representation (spatial mental model) of fairly elaborate spatial descriptions. Secondly, we examine whether blind people exhibit the same preferences (i.e. level of performance on spatial tasks) as sighted people in processing the type of perspective that is employed in a spatial description. Early blind, late blind and sighted participants listened to a route and a survey description of two environments. Next, they had to execute a recognition/priming task, a bird flight distance comparison task, and a scale model task. Spatial priming and symbolic distance effects were found for all participants. These findings suggest that early and late blind people can form spatial mental models on the basis of route and survey descriptions. Interestingly, in contrast with sighted people, blind people performed better after listening to a route than a survey description, even when the spatial problems that has to be solved explicitly favor the survey description. It seems that people with active vision build up a spatial mental model more efficiently from a survey description, while people with only visual memories (late blind), similar to people with no visual memories (early blind), build up a spatial mental model more efficiently from a route description.