Climate change is a global threat to species, and their capability to invade and colonise new landscapes could be limited by the habitat fragmentation. Improving landscapes by adding additional resources to landscapes is an important initiative to restore habitats. Such improvements will be particularly important to promote species recovery in fragmented landscapes and to understand as well as facilitate range-shifting for species (also called an invasion). We use a recent method to approximate the time taken by species to invade landscapes and reach the new areas of suitable environment, which based on network flow theory. Based on this, we propose and test a new method that can help to compute the best locations in landscapes in order to restore habitat which leads to minimising the expected time taken by species to invade and reach targets. The new optimisation method has been compared with other two baseline methods. The evaluation conducted using real heterogeneous landscapes shows that the proposed method outperforms the competitive baseline methods in terms of proposing landscape modifications that minimise the expected time of the invasion process.