The integration of intersecting routes is an important process for the formation of cognitive maps and thus successful navigation. Here we present a novel task to study route integration and the effects that landmark information and cognitive ageing have on this process. We created two virtual environments, each comprising five places and one central intersection but with different landmark settings: in the Identical Landmark environment, the intersection contained visually monotonic features whereas the intersection contained visually distinctive features in the Different Landmarks environment. In both environments young and older participants were presented with two short routes that both traversed through the shared intersection. To test route integration, participants were asked to either repeat the learning routes, to navigate the routes from the destination to the starting place or to plan novel routes. As expected, results demonstrate better performance when repeating or retracing routes than when planning novel routes. Performance was better in younger than older participants and in the Different Landmark environment which does not require detailed knowledge of the spatial configuration of all places in the environment. A subgroup of the older participants who performed lower on a screening test for cognitive impairments could not successfully complete the experiment or did not reach the required performance criterion. These results demonstrate that strategically placed landmarks support the integration of route knowledge into spatial representations that allow for goal-dependent flexible navigation behaviour and that earliest signs of atypical cognitive ageing affect this process of route integration.