Effective detection of islanding and rapid distributed generation (DG) disconnection is essential to prevent safety problems and equipment damage in power systems. The common islanding protection method is based on passive techniques that do not perturb the system but have large non-detection zones. Therefore, the first part of this thesis attempts to develop a simple and effective passive islanding detection method with reference to a probabilistic neural network-based classifier, as well as utilizes the features extracted from three-phase voltages seen at the DG terminal. This approach enables initial features to be obtained using the phase-space technique. Meanwhile, the second part of the thesis focuses on the development of an optimal load shedding scheme after the system experiences an unintentional islanding state to prevent system collapse. A novel heuristic technique based on the backtracking search algorithm is proposed as an optimization tool for determining the optimum load shedding based on the proposed objective function.