The realisation of the Smart Grid envisions end-users, equipped with renewable energy generators such as solar panels, to act as “prosumer”. The prosumer is engaged in both energy production and consumption, and its energy can be exchanged as a commodity between end-users, disrupting the traditional power system distribution model. To facilitate the transition to a prosumers’ governed grid, we proposed a novel strategy for optimising decentralised energy exchange in the Smart Grid and tested it on a radial test feeder. In this paper, we consider the topological effects on the distribution of evaluating various topologies of the test feeder based on the optimisation strategy proposed in our previous work. We simulate the energy exchange based on the Monte Carlo method and compare the performance of different topological shapes. Overall, the complete graph has better performance than other topological shapes. But it is not a practical topology because of its high implementation costs. For other topologies, the random graph has higher efficiency than others in reducing the maximum load in electric lines and hopping count of energy delivery, and the small-world is relative more efficient in decreasing energy loss of delivery and energy costs for end-users.