Ferroresonance, a non-linear and unpredictable disturbance, is rare compared to traditional power system faults occurring in power systems. This rarity, coupled with its complexity, makes it a challenging phenomenon to be detected and identified. This work presents a detection and classification scheme for ferroresonance and its modes. It is carried out by continuously processing the three-phase voltage and current signals using the discrete wavelet transform (DWT). The developed models are simulated in electromagnetic transient software and processed using the DWT to extract fault signatures and predictors. A decision tree classifier is trained to detect and classify a disturbance as ferroresonance using an adaptive time based on the disturbance class. The computational burden of the detection and classification process is significantly reduced by using the superimposed component of the voltage and current to detect transient inceptions before classification. Furthermore, the classification of different modes and classification from other non-linear faults, such as arcing faults, is discussed. The adaptive timing and detection scheme demonstrates that the proposed methodology is efficient and can classify the disturbance into different modes.