Circulation network design, widely mentioned as road network design, is one of the main topics in urban planning, urban transportation planning, traffic engineering and urban design. In the literature on road network design, problems and principles are examined through quantitative and qualitative researches (Magnanti and Wong, 1984; Hillier, 1987; Ran and Boyce, 1996; Hillier, 1999; Southworth and Ben-Joseph, 2003; Drezner and Wesolowsky, 2003, Marshall, 2005). The former one implies analysis and simulation of demographic and economic variables in order to apply appropriate circulation schemes, while the latter one deals with various factors of design and aims to create urban spaces regarding ‘space quality’. This paper examines the role of urban patterns in the formation of road networks; thereby depicting relative performance and handicaps of road networks under stated travel distribution among specified units. Road network design requires knowledge of circulation concepts and methodology of assessment. This study provides a topological analysis method and assessment criteria in order to understand performances of various types of road networks. In order to constitute road networks, circulation systems are adapted on three basic types of urban pattern: the grid, the hexagon and the radio-centric. Finally, two certain travel demand matrices are assigned to the specified road networks in order to assess relative advantages of various patterns and circulation networks. At the first stage, travel distribution among units (zones) is assumed to be uniform while second stage is performed with a matrix representing a centre-oriented travel distribution pattern. Here, circulation systems connote to the conceptual representation of movement of vehicles on specific routes, while circulation network (road network) is the concrete form of any specificied circulation system. There are numerous studies on advanced traffic management systems, which mainly focus on the issue of uninterrupted traffic flow (Ran and Boyce, 1996; Levinson, 2003). In the last decade, intelligent transportation systems and traveler route guidance stystems are extensively adapted to urban transportation systems. On the other hand, there are limited studies measuring the the performances of such systems, which are highly dependent to the characteristics of road network (Emmerink et al., 2003; Levinson, 2003). Marshall (2005) argues that street pattern is significantly influenced by the “geometry of movement” and “the topology of route connectivity”. Geometry of movement is the representation of individuals’ travel patterns while connectivity refers to the structural (topological) characteristics of road network; such as hierarchy and composition of links (road sections) and nodes (junctions). This study deals with the role of urban patterns on the formation of road networks and relative performances of circulation systems. Conceptual forms of circulation systems, described by Wright et al.(1989) are adapted to basic types of urban patterns (the grid, the hexagonal and the radiocentric) in order to obtain road networks. Those networks are compared in terms of network efficiency, measured by average travel distance, traffic volume/capacity ratio, standard deviation among travel volumes, and average travel time. Proposed measures are applied to 33 test networks, which are derived from three base patterns: the grid, the hexagon, and the radio-centric. Traffic assignment procedure is carried out in order to observe performances of road networks. Results showed that when uniform travel distribution is considered, connected networks, such as convex polygons and concentric corridors, perform better than less connected circulation networks. Those networks perform with uniform volume capacity ratios; therefore, hierarchial capacity design is not necessary for such systems. Convex systems have high efficiency when adapted to grid patterns rather than hexagonal and radio-centric. Average traffic volumes and standard deviation on convex networks are much less than semi-connected or tree networks. Hence, connected road networks perform with less average travel distance and average travel time. Results showed that grid patterns are more advantageous when convex circulation systems are adapted, while hexagonal patterns perform better when tree or star shaped systems are adapted. Compared to two-way systems, one-way circulation systems prove to be inefficient in terms of travel distance and time, however, conflicts at intersections are much less. When centre-oriented trips are assigned to road networks, star-shaped and tree-shaped networks perform better, however, total travel time and total travel distance is more than others. Those networks perform with moderate volume capacity ratios.