Abstract: Thermal conductivity is an essential parameter for studying coupled transport of heat and water in soil. In order to determine the error of multi-needle heat pulse technique, measurement accuracy of a multi-needle heat pulse probe was tested by comparing its data with that from the single probe method in this paper. Thermal conductivity measurements were conducted on four soils of different textures under various bulk densities, water contents, and air pressures. The results indicated that generally the thermal conductivity from the multi-needle heat pulse method agreed well with that from the single probe method, and the method was able to provide accurate soil thermal conductivity data. The mean error of the multi-needle heat pulse probe thermal conductivity was 0.074 W/(m·K). Due to the mean free path of gas decreased, the thermal conductivity of dry soil showed a logarithmic increasing with the increasing of air pressure. The errors of the multi-needle heat pulse probe mainly appeared at the medium water contents, where larger temperature rise near the heater enhanced latent heat transfer of vapor. The contact resistance between the needle and sample, change of soil bulk density near the needles, and liquid water transfer under temperature gradient may also have contributed to the errors in the muti-needle probe method. This study will provide a method for estimating soil thermal conductivity in the soil and water engineering.