This paper deals with charging optimisation of an electric vehicle (EV) fleet, which is aimed at real-time applications which includes both charging cost and peak power minimisation. It is shown that the optimal charging problem which includes these two criteria can be modelled as a quadratic optimisation problem, and can, thus, be solved with widely available and computationally efficient quadratic programming solvers. The corresponding quadratic cost function is extended with a single, free parameter, in order to enable tuning of relative significance of charging cost- and peak power-related sub-cost functions. The functionality of the proposed charging optimisation approach is demonstrated for the case of an isolated electrical energy system and hypothetical EV fleet of the National park Mljet. The time distributions, needed as inputs for charging optimisation, are obtained by using appropriate EV models and available driving cycles and related schedules of the particular fleet. Finally, benefits of using the presented charging optimisation method are presented based on comparison with respect to uncontrolled charging (charge when connected) results in terms of charging cost and peak power values.