Although multimodality has been widely studied in the literature, planning and operating bus lanes in congested urban city centers are still challenging topics for researchers and policy makers. Most existing approaches lack quantitative methods for estimating the impact of bus lanes or for optimizing the operation of bus lanes at a system level. This paper proposes a novel optimization approach for allocating road space to bus lanes in cities. The approach determines the optimal space share between the modes in service and allocates the bus lanes by integrating strategies that lead to less total travel cost. By relying on recent advances in network-level traffic flow modeling, namely, the multimodal macroscopic fundamental diagram (mMFD), the approach captures multimodal traffic dynamics and travel costs by mode. The impact of a bus lane on mode usage is taken into account to aggregated mode shift phenomena under changes in layout of dedicated bus lanes. Simulation was performed in a Swiss city network to test the proposed optimization approach. The research found that (a) the mMFD could be properly integrated to decide for road space optimization of large-scale multimodal urban networks, (b) an optimal and efficient space share minimized the total travel cost for all users, and (c) the best strategy for the studied network was to implement the allocated space on the connected links on a corridor rather than to assign them sparsely to the links that are heavily congested.