The use of soil heat flux as a critical component of the surface energy balance is routine; however, its accurate quantification is not. The direct measurement of soil heat flux is generally accomplished with soil heat flux plates. However, the presence of heat flux plates causes perturbations in heat and fluid flow in the soil that may give rise to measurement errors. We describe here the direct measurement of soil heat flux with the gradient technique using a tri-needle instrument, where soil heat flux is the product of the soil thermal conductivity measured by transient heating of the center needle and the soil temperature gradient measured between the outer needles. Soil heat fluxes measured this way were compared with those obtained from commercially available soil heat flux plates. Laboratory trials revealed good precision and accuracy (error generally <5%) in measurements of thermal conductivity and soil heat flux with the heated needle technique in a medium-textured sand, but exposed occasional errors in the accurate determination of thermal conductivity arising from poor sensor-soil thermal contact in a coarser medium. Extended field data in a finer-textured soil showed good agreement among needle sensors and heat flux plates, and surprisingly did not reveal errors associated with fluid blockage by the heat flux plates. Our results indicate that the measurement of soil heat flux with the gradient technique with needle sensors is a viable alternative to heat flux plates, and may improve absolute accuracy of this measurement.