The transformation of urban roadways into pedestrian streets is a popular measure for reshaping city parts and enhancing their livability. Nevertheless, pedestrianization schemes are expected to have some impact on the performance of the neighboring road network, especially if these are established ad-hoc or solely based on non-transport criteria. This study introduces a methodological tool for supporting decisions on implementing pedestrianization schemes in urban networks. A bi-level network design model variant is developed for that purpose, whose design objective is to maximize the extent of pedestrian streets in an urban network, while maintaining acceptable impacts to the performance of the road network. Alternative decisions on pedestrianization are considered for each network segment; these include partial (one-directional) or complete (bi-directional) pedestrianization under physical and operational criteria and constraints. The model is applied for a mid-sized urban network in Greece and solved using a genetic algorithm. Results show that the pedestrianization of almost 7% of the road network in relation to length leads to a 40% increase in total network travel time, while a higher ratio of complete versus partial pedestrianization is more advantageous. Outcomes also reveal that that rigid design guidelines should be examined in a case-by-case approach, as superior results may be attained if some constraints, such as those related to the overall street width, are relaxed. Reasonably, policy priorities significantly impact generated solutions and are expected to play a decisive role in the design of pedestrianization schemes.