Implementation of distributed generator (DG)-based network in modern power system is a mandatory move after one decade of deregulated electrical power industry in India. The protection scheme has to be installed to protect this DG-based system in undesired situation of frequent transient and steady state disturbances to supply the uninterrupted and good quality power. To do this task, a smart and reliable protection scheme which can detect and classify the islanding condition, fault types, and steady state operation is designed in this work. The disturbance signals are generated from a wind and solar photovoltaic (PV) generation-based hybrid distribution network. This scheme is a combination of two different wavelet-aided machine learning (ML) classifiers. The input datasets for these classifiers were built using feature selection based on principal component analysis (PCA). The performances of classifiers have also been evaluated by means of confusion matrix. Every class of disturbances can be detected accurately in fast, and the proposed algorithm has an added advantage of incorporating large number of disturbances and more classifiers in future.