The purpose of this study is to explore human wayfinding behavior within a familiar street network in a purely qualitative manner. So, the most significant criteria of making decision in the network is the cognitive map which has been learnt before during the human beings daily activity. The cognitive map includes a great deal of details of environmental descriptions but it seems that preacquainted people find routes in the street network by finding a connection from an origin to skeleton of major paths then moving to the proximity of the destination. Hence, we evaluate a heuristic approach for qualitative wayfinding proposed by Kuipers et al. (2003) named boundary relations. In the process of implementation we encountered some problems which guide us to some considerations for refinements and developments. The concept of cardinal direction is added to the properties of paths of the skeleton and we use the qualitative spatial reasoning by means of it in the sub-goal refinement stage. In addition, we introduced a new obvious parameter that is used as an input of the problem; it is the cardinal direction from origin to destination. As a practical activity, the Dijkstra algorithm and the developed heuristic are implemented in the street network of Tehran, the capital of Iran, which has a complicated traffic problem. The outcomes of these algorithms are depicted in this paper that the first one is considered as an optimum way and the second one is a simulation of human behavior.