We conducted time-lapse resistivity imaging during pumping tests at the Pengtsuo and Dajou test sites in Taiwan in order to examine the feasibility of estimating hydrogeological parameters with resistivity variations. Core logs reveal that the subsurface consists mainly of at least 100-m-thick gravel and sand at the two test sites. The resistivity differences between the pumping stages and pre-pumping background are well correlated to water level changes that are due to the dewatering of pumping activity. Therefore, it is possible to use the geometry of resistivity anomalies to estimate the hydraulic conductivity of the unconfined aquifer using the distance-drawdown equation for pumping tests in unsaturated aquifers. For each site, we used the contours of resistivity variations and recorded water levels in the pumping well to depict the bottom of the drawdown cone. The estimated hydraulic conductivity and specific yield, respectively, are 1.33 × 10 − 4 m/s and 0.12 at the Pengtsuo site, and are 2.50 × 10 − 4 m/s and 0.22 at the Dajou site. These values are consistent with the parameters that engineers from Taiwan Sugar Company calculated previously regarding groundwater-level variations in multiple wells (9.65 × 10 − 5 m/s and 0.13 at Pengtsuo, and 1.00 × 10 − 3 m/s and 0.19 at Dajou). This consistency suggests that resistivity imaging can perhaps serve as an alternative way to yield information about hydrogeological parameters.