We deployed a dense seismic array to image the shallow structure in the injection area of the Brady Hot Springs geothermal power plant in western Nevada. The array was composed of 238 three‐component, 5 Hz nodal instruments, 8700 m of distributed acoustic sensing (DAS) fiber‐optic cable (FOC) installed horizontally in surface trenches, and 400 m of FOC installed vertically in a borehole. The geophone array had about 60 m instrument spacing in the target zone, whereas DAS channel separations were about 1 m with an averaging (gauge) length of 10 m. The acquisition systems provided 15 days of continuous records, including active‐source and ambient noise signals. A large vibroseis truck was operated at 196 locations, exciting a swept‐frequency signal from 5 to 80 Hz over 20 s using three vibration modes (vertical, longitudinal, and transverse), with three sweeps per mode at each site. Sweeps were repeated up to four times at each site during four different stages of power plant operation: normal operation, shutdown, high and oscillatory injection and production, and normal operation. After removal of the sweep signal from the raw data, the first P‐wave arrivals were automatically picked using a combination of methods. The travel times were then used to invert for the 3D P‐wave velocity structure. Models with 100 m horizontal and 20–50 m vertical node spacing were obtained, covering an area 2000 m by 1300 m, with acceptable resolution extending to about 250 m below surface. The travel‐time data were fit to a root mean square (rms) misfit of 31 ms, close to our estimated picking uncertainty. Lateral boundaries between high and low velocity zones agree relatively well with the location of local faults from previous studies, and low near‐surface velocities are associated with faults and fumarole locations. A sharp increase in velocity from to at approximately 50 m below the ground surface in many parts of the study area may indicate a shallower water table than expected for the region.