Disturbance to the output of a linear voltage regulator (LVR) due to an abrupt change in output current can be compensated by using an output capacitor. A capacitor has an internal parasitic resistive element known as equivalent series resistance (ESR). However, the ESR value may change due to aging and temperature variations, forming a failure region in an LVR. Besides, the performance of each manufactured LVR varies due to differences in the manufacturing process. Consequently, failure region determination (FRD) for LVRs involves time-consuming and costly manual data acquisition and requires analysis to determine the failure region accurately. In this work, efficient and effective FRD methods were developed by applying the concept of virtual sensing. Two approaches were used, namely, a data-driven approach (DDA) and a model-based approach (MBA). The developed FRD methods were as follows: the data interpolation method (DDA-DIM), the input-output model-based method (MBA-IOM), and the circuit analysis model-based method (MBA-CAMM). DDA-DIM utilizes a multilayer perceptron and a radial basis function neural network. Meanwhile, MBA-IOM and MBA-CAMM estimate the black-box model of an LVR and the circuit analysis model, respectively. The results of the three methods were compared with the benchmark developed using manual FRD. MBA-CAMM was determined as the most effective and efficient FRD method that applies the virtual sensing concept.