Organic Rankine Cycle converts low grade heat sources into power utilizing organic fluids with low boiling temperature and pressure. In this cycle the design and performance of the expander has a significant impact on the cycle's overall efficiency. This work presents an integrated mathematical approach for the development of an efficient and compact small-scale radial turbine. This mathematical approach integrates the mean-line modelling with real gas formulation and GA(genetic algorithm) optimisation technique. In this methodology, the mean-line modelling coupled with real gas formulation is employed to perform parametric studies to identify the key variables that have significant effect on the turbine efficiency. Such variables are then used in the GA to optimise the turbine performance. Eight organic fluids are investigated to optimise the performance of the small-scale radial turbine in terms of efficiency. Results showed that the achieved radial turbine efficiencies vary from 82.9% to 84%; which is higher than the reported efficiency values of other types of expanders. R152a showed the highest efficiency of 84% with seven degrees (K) of superheating. However, if the superheating is to be avoided, isobutane exhibited the most favourable characteristics in terms of efficiency (83.82%), rotor size (66.3 mm) and inlet temperature (89.2 degrees C). (C) 2015 Elsevier Ltd. All rights reserved.