Abstract: The Beijing-Tianjin-Hebei (BTH) region serves as a critical area for comprehensive groundwater overexploitation management in North China. Analyzing variations in groundwater levels and storage, as along with their driving factors, is of great significance for groundwater management. This study systematically investigated the regional groundwater dynamics using groundwater levels, GRACE satellite observations, precipitation, and extraction data, through water balance analysis, Mann-Kendall trend tests, correlation analysis, and cross-wavelet transform methods. The main conclusions are as follows: 1) Groundwater levels in the BTH plain showed continuous decline before 2014 but significant recovery during 2015-2022. In Beijing Plain, the declining rate of shallow groundwater (−0.85 m/yr) was greater than that of deep groundwater, whereas in the Tianjin and Hebei Plains, the decline rate of shallow groundwater levels was lower than that of deep groundwater levels. The shallow groundwater level in Beijing Plain changed from declining at 0.85 m/yr to rising at 1.1 m/yr. In Tianjin, the groundwater level at the depression cone center of the third confined aquifer shifted from declining at 0.28 m/yr to rising at 2.8 m/yr. In Hebei Province, the declining rate of shallow groundwater decreased from 0.28 m/yr to 0.17 m/yr, with rapid recovery beginning in 2021. This recovery trend mainly resulted from the combined effects of increased precipitation, groundwater extraction restriction policies, and ecological water replenishment projects. 2) Groundwater storage showed relatively slow decline during 2003-2007 but accelerated after 2008, with the most extensive depletion occurring in summer and autumn. 3) Groundwater storage and level variations exhibited high consistency. In Beijing, the correlation between shallow groundwater storage and level changes (0.6) was higher than that for deep groundwater (0.53), while in Tianjin (0.76 vs 0.40) and Hebei (represented by Hengshui: 0.45 vs −0.28), deep groundwater showed stronger correlations. 4) Cross-wavelet analysis revealed significant annual-scale resonance between groundwater level/storage and climatic factors, with 4.5-6 months lags in response to precipitation and approximately 3-month lags for evapotranspiration.