Rivers of the northern Great Plains have lacked long-term, continuous water temperature assessments, and there is limited information on thermal regimes of these systems and factors driving water temperature. We collected and assembled 2001–2022 water temperature data from 18 sites on four reaches of three rivers that differ in anthropogenic impacts: semi-natural Yellowstone River (YR), flow-impacted Milk River (MK), 351-km of the Missouri River affected by hypolimnetic releases from Fort Peck Dam (FPD), and the semi-natural Missouri River (MR3093) upstream from FPD. Objectives were to: (1) compare May–September mean daily water temperature (T_w), day of year of maximum water temperature (T_maxdoy), and maximum water temperature (T_wmax) among reaches, (2) evaluate air temperature (T_a), river discharge (Q_w), and dam-release water temperature (T_wdam) as T_w drivers, and (3) model longitudinal recovery of T_w downstream from FPD. Mean T_w and T_wmax were greatest at the YR, MR3093 and MK sites, and significantly less through 291-km downstream from FPD. T_maxdoy at initial sites downstream from FPD was delayed 43–69 days relative to the semi-natural reach upstream from FPD. T_a was the primary correlate of T_w for the semi-natural sites; whereas, T_wdam and T_a varied inversely as primary drivers for sites downstream from FPD. Thermal recovery from hypolimnetic releases was incomplete 291-km downstream from FPD and warming persisted 351-km downstream. Results quantify the varied water temperature regimes of rivers in the northern Great Plains and improve understanding of controls affecting T_w among reaches. Water temperature attributes of semi-natural reaches could be used as restoration targets for 300-km of Missouri River presently impacted by hypolimnetic releases.