The microgrid technology must be conforming in electrical power system to cater the expanding development in power demand in the present scenario, and it is extremely desirable to maintain power quality of those networks in competitive environment. In such a situation, the identification of transient disturbances and classification are always exigent because of the power quality issues. This work presents a novel decision tree (DT) algorithm based on machine learning for the identification and classification of different possible transient conditions in microgrid. The transient study comprises of fault at distribution line, i.e., symmetrical and unsymmetrical, islanding from grid due to three-phase short-circuit fault, intentional islanding, etc. The proposed method starts with tracking the transient signal of phase voltage and current at different locations, and each signal is to be processed through discrete wavelet transform (DWT). The algorithm is followed by the feature selection from the performance indices of DWT coefficient for DT classification technique. The findings of the extensive case studies show that the suggested algorithm is used to detect and discriminate each transient event with 100% accuracy. The performance of the proposed method has been validated using the confusion matrix.