The thermal conductivity probe system has been used in food. However, since the probe system is derived from an idealized transient heat transfer model, there are inevitable differences between the real probe and the theoretical model. The test error resulted by some experimental operating parameters including input voltage, heating time of thermal probe and size of sample was studied. The best values of input voltage, heating time of thermal probe and size of sample were obtained to minimize the test error. The optimal value for input voltage of thermal probe is in the range of 2.5～6.5V, and the optimal heating time of thermal probe is in the range of 20～50s. And then the thermal conductivities of tomato fruits at different ripeness stage were measured. The experimental data can give a reference for storage and process of tomato industry.