In this paper we study local routing strategies on geometric graphs. Such strategies use geometric properties of the graph like the coordinates of the current and target nodes to route. Specifically, we study routing strategies in the presence of constraints which are obstacles that edges of the graph are not allowed to cross. Let P be a set of n points in the plane and let S be a set of line segments whose endpoints are in P, with no two line segments intersecting properly. We present the first deterministic 1-local O(1)-memory routing algorithm that is guaranteed to find a path between two vertices in the visibility graph of P with respect to a set of constraints S. The strategy never looks beyond the direct neighbors of the current node and does not store more than O(1)-information to reach the target.We then turn our attention to finding competitive routing strategies. We show that when routing on any triangulation T of P such that S⊆T, no o(n)-competitive routing algorithm exists when the routing strategy restricts its attention to the triangles intersected by the line segment from the source to the target (a technique commonly used in the unconstrained setting). Finally, we provide an O(n)-competitive deterministic 1-local O(1)-memory routing algorithm on any such T, which is optimal in the worst case, given the lower bound.