In this work, empirical relationships based on indices (AE, IL, and KP), an empirical model (Weimer-2005), coupled space weather models (BATS-R-US+ CRCM/RCM), and combinative methods (using data from SuperDARN for electric field and OVATION PRIME+ IRI combination for the conductances) are employed to investigate the Joule heating (JH) variability during two isolated substorms on 8 March 2008. The results are compared according to substorm phases, magnetic local time sectors, magnetic latitudes, role of electric fields, ionospheric conductances, and neutral wind. Qualitatively, all methods exhibit similar variations in the estimated JH rates. AE-dependent methods and the Weimer-2005 model show the highest JH rates. The response of AE to the elevated Epsilon levels is found to be delayed up to 50 min, whereas the onsets and variations in the other methods are concurrent with Epsilon. The combination of OVATION Prime+ IRI and SuperDARN provides values close to the (Foster et al., 1983, https://doi. org/10.1029/JA088iA06p04885) suggesting that it is representative enough to use for estimating the JH. The OVATION Prime+ IRI and SuperDARN method can provide JH at any point in the solar cycle based on dynamic solar wind input. BATSRUS/CRCM conductances and electric fields are higher than BATSRUS/RCM leading to 2.5 times as much JH as estimated by BATSRUS/RCM on average. Neutral winds are found to comprise an overall 15% of total JH with a maximum of 25% in dusk sector during the expansion phase. We discuss that W05 and AE-dependent methods overestimate the JH rates during the solar minimum since they are constructed using solar maximum data.