Abstract: In recent years several European cities have introduced road pricing as a tool for managing transport demand, especially to reduce traffic congestion and rebalance the modal split between private vehicles and mass-transit systems. Indeed, user behaviour brings about a User Equilibrium condition, which does not correspond to overall utility maximisation and fails to take account of external costs. Hence, in order to achieve the efficient use of transportation systems (System Equilibrium), tolls can be charged on urban roads so that the social surplus is maximised. For several reasons (theoretical, political, social acceptability) it is impossible to charge efficient tolls (first-best solutions) proposed in the literature; therefore in real networks sub-optimal tolls (second-best solutions) are applied. In this study we analyse the effects on optimal fare design when pricing revenues are wholly or partly used for improving public transport. In particular, we formulate a model according to economic theory in a multimodal and multiuser context, where multimodal features are calculated explicitly on the network for each fare configuration. The model is applied on a trial network (built with heterogeneous values of relative accessibility among different traffic zones) and several second-best strategies are analysed with particular attention to the use of pricing revenue.