Ocean heat content (OHC) is a fundamental indicator of Earth's energy imbalance and accurate OHC estimates are essential for monitoring climate variation. Reliable estimates of historical OHC changes are essential for reducing observational uncertainties and improving confidence in understanding and predicting decadal climate variability. Various ocean gridded products have revealed continuous upper-ocean warming. However, the consistency of OHC changes across different datasets and time periods remains uncertain. Here, we compare OHC (0–700 m) changes during the satellite era (1993–2023) and Argo era (2005–2023) using multiple ocean reanalysis (RAs) and objective analysis products (OAs). Results show that OAs and RAs demonstrate remarkable consistency in characterizing the mean state and variability of global OHC during the satellite era. Particularly, OAs exhibit a linear warming rate of 6.3 ± 0.6 ZJ yr⁻¹, closely matching the RA ensemble mean (6.6 ± 3.2 ZJ yr⁻¹), although the latter exhibits substantially larger uncertainty. This larger spread among RAs is observed across all ocean basins, especially within highly dynamic regions, and persists into the Argo era. Critically, the OHC trend exhibits excellent latitudinal consistency in both OAs and RAs, with peak warming along 35–39°N and 45–41°S. While the strongest warming during the satellite era occurs in the Southern Hemisphere, the Argo era reveals pronounced acceleration in the Northern Hemisphere, with the Pacific and Atlantic Oceans contributing ~60% and ~40%, respectively. Ocean heat budget analysis based on OAs reveals that ocean heat uptake consistently contributes to ocean warming in both eras, while ocean dynamic processes shift from suppressing to amplifying OHC increase in the Argo era (especially post-2018), driving accelerated Northern Hemisphere warming. However, this mechanism is not captured by most RAs, highlighting the importance of reducing uncertainties in ocean reanalysis to improve our understanding and constrain estimates of decadal climate variability.