The effects of the injection parameters on delivery of nanoscale zero-valent iron (NZVI) to contaminated groundwater were investigated. The first two NZVI injections (gravity injection at low flow rates) resulted in NZVI being poorly mobile and gave total cumulative mass recoveries at the monitoring wells of 1.07%–2.43%. NZVI reached some wells (KDMW-3, MW-2, MW-4, and MW-7) earlier than the bromide tracer. The dominant travel directions for NZVI and the bromide tracer were very different. The NZVI transport characteristics suggested that targeted NZVI delivery requires preferential groundwater flow paths and local heterogeneity to be considered. In the gravity injection tests, the maximum NZVI concentrations and cumulative NZVI mass recoveries in the wells decreased markedly as the injected NZVI concentration and dose increased. In the third and fourth tests, in which NZVI was injected under pressure at high flow rates, NZVI was effectively delivered to the wells despite the injected NZVI concentration and dose being high. Relatively high cumulative mass recoveries of 26.0% and 74.5% were found for the third and fourth injections, respectively. Controlling the flow rate (pressure) and NZVI concentration and dose simply and effectively controlled NZVI mobility in the groundwater. The colloidal and electrostatic characteristics of the NZVI particles were monitored and modeled, and the results indicated that NZVI particles without Derjaguin–Landau–Verwey–Overbeek energy barriers were successfully delivered to the target zone and that decreased magnetic attractive forces between NZVI particles caused by iron corrosion probably decreased the degree of NZVI particle aggregation and therefore contributed to NZVI being delivered to the target zone.