As an energy storage or power buffering device, supercapacitor (SC) has incomparable advantages in terms of charging-discharging efficiency, response speed, and durability. Generally, a simplified energy estimation method based on a constant or rated capacitance can be used to evaluate the state-of-energy (SoE) of SCs. However, using a constant capacitance to evaluate the SoE of SC is very imprecise. Indeed, the SC internal parameters would change in accordance with the operating current and voltage such that the actual capacitance is far from the rated capacitance. This letter presents: a variable first-order resistance-capacitance equivalent circuit for dynamic modeling of the SC; a real-time voltage state observer used to estimate the variable internal capacitance and resistances of the SC; three different energy estimation methods; and simulated and experimental validations of the adaptive energy estimation method based on a 3000 F SC in electric vehicle applications. By using the SoE based on an experiment calibration as a reference standard, the proposed adaptive energy estimation method can improve over 0.75% of the estimated SoE error for the SC compared with the conventional estimation method.