Small-scale Organic Rankine Cycles (ORCs) represent a promising technology for on-board waste heat recovery (WHR) from internal combustion engines (ICEs) especially in innovative long-haul trucks transport applications. Despite market leader companies have already proved its effectiveness, detailed system design procedures are scarcely available in the open literature and proposed solutions are often more simplified with respect to current industrial state-of-art. The present work describes a methodology to include within the ORC design and optimization procedure an efficiency map of the turbo-expander, retrieved exploiting a reduced-order method developed in-house. The proposed approach allows therefore to design the thermodynamic cycle considering a realistic performance of the expander and to retrieve the best cycle architectures and turbine geometry depending on heat source characteristics and active constraints. First the turbine performance is defined in general terms and investigation of its applicability to different fluids and to expansions characterized by various degrees of non-ideality is provided. Then, the methodology is applied to different WHR applications highlighting for several potential working fluids the impact of adopting turbine performance maps within the cycle optimization procedure rather than assuming constant efficiency values. Results show that the cycle configuration and the main design parameters are strongly affected by the plant size and the working fluid selection, and that the adoption of a turbine performance map is particularly valuable when dealing with small size power plants adopting high density fluids. In particular, the assumption of constant turbine efficiency may lead to the definition of thermodynamic cycles extremely far from the optimal ones (e.g. adoption of supercritical cycles rather than subcritical ones) that in turn can result in an unfeasible design of the turboexpander. Moreover, this assumption may also lead to an incorrect comparison between different candidate working fluids lacking in identify those more suitable for a specific plant size. Both those limits can be solved with the adoption of the expander performance map proposed in this work. Final results show that on board WHR ORC can reach an efficiency of at least 20% in a relatively large range of internal combustion engine size. On the contrary, the choice of the best working fluid is affected by engine size and chosen hot sources: for small available thermal power the use of complex and flammable hydrocarbons (cyclopentane, cyclohexane, toluene) is the most efficient choice while the use of low GWP non-flammable halogenated hydrocarbons (R1233zd and R245fa) becomes promising only for relatively large available thermal power.