AbstractThe power energy input carried by precipitating electrons into the auroral zone is an important parameter for understanding the solar wind‐magnetosphere energy transfer processes and magnetic storms triggering. Some magnetic storms present a peculiar long recovery phase, lasting for many days or even weeks, which can be associated with the intense and long‐duration auroral activity named HILDCAA (High Intensity Long Duration Continuous AE Activity). The auroral energy input during HILDCAAs has been pointed out as an essential key issue, although there have been very few quantitative studies on this topic. In the present work, we have estimated the auroral electron precipitating energy during the events of long (LRP) and short (SRP) storm recovery phase. The energy has been calculated from the images produced by the Ultraviolet Imager (UVI) on board the Polar satellite. In order to obtain accurate energy values, we developed a dayglow estimate method to remove solar contamination from the UVI images, before calculating the energy. We compared the UVI estimate to the Hemispheric Power (HP), to the empirical power obtained from the AE index, and to the solar wind input power. Our results showed that the UVI electron precipitating power for the LRP events presented a quasiperiodic fluctuation, which has been confirmed by the other estimates. We found that the LRP events are a consequence of a directly driven system, where there is no long‐term energy storage in the magnetosphere, and the auroral electrojets during these events are directly affected by the electron precipitating power.