【Objective】The use of resistivity method in soil test has a number of shortcomings in accuracy and electrode is an important factor affecting accuracy of the test. In most tests, efforts are made to minimize contact and select appropriate electrode distance on the simple operational level. So accuracy of the test results is hard to correct and unify. Therefore, systematic analysis and quantification of these problems are required. If it is feasible to determine influences of contact and distance of the electrodes on test results, interference of the electrode can be avoided to the maximum extent, especially in pole-pole method. It is expected that this study may provide certain reasonable suggestions for electrode placement to minimize deviation of the test results in using the resistivity test device.【Method】In this paper, the soil-box resistivity test model was used to measure soil resistivity in zinc-contaminated sand used as simulated soil in the test. The test was designed to have three variables, i.e. soil water content, diameter of the electrode and distance between two electrodes for measurement of its AC resistance and resistivity (simply referred to as resistance and resistivity in the forthcoming paragraphs).Through the test with the indoor system and simulation of COMSOL physical field, influences of electrode contact degree and distance between electrodes on resistivity and its measurement were explored, a corresponding resistivity accuracy analysis model established, influences of contact and distance of electrodes on the test results discussed, and comparative analysis of errors and deviations of the test performed. 【Result】Results show that the experimental research and numerical simulation could verify and supplement each other, which indicates that the electric conductivity test model derived in this paper is correct. The electric conduction model could be divided into a normal section and transition sections at each end, connected in tandem. By means of the simulation, radial current density is used to reflect lengths of the transition sections and figured out the degree of current deviation. The change of current flow network produced shrinkage resistance in transition sections. Both shrinkage resistance and lengths of the transition sections decrease with the increase of contact degree, but are not affected by the distance between electrodes. After the length of the transition section was worked out, deviations of the resistivity of the transition and normal sections could be calculated numerically and accurately in line with the equation for calculation. Error analysis shows that the measured value of comprehensive resistivity gradually approaches that of normal resistivity with the expanding contact area and the extending distance of the electrodes. 【Conclusion】In summary, to keep electrodes in contact and properly apart is an effective way to reduce relative errors of the test to measure resistivity. In this study through numerical analysis, it is leant that deviation of the measurement of resistivity is caused by incomplete electrode contact and resistivity values of the normal sections can be worked out. All the findings in this study may serve as theoretical guidance and technical support for perfect electrodes arrangement and higher accuracy of the measurement of resistivity.