The most common direct and indirect measures of soil compaction are bulk density and soil mechanical resistance, respectively. The cone penetrometer is used traditionally for measuring soil mechanical resistance. However, the use of this device is time-consuming and the measurement is made at a single discrete point; therefore, it is not suitable for large-size farms. The multiple-tip sensors are used for continuous measurement of horizontal soil mechanical resistance in order to produce spatial mapping of soil compaction. The objective of this study was to improve the performance of an on-the-go horizontally-operated three prismatic tips penetrometer developed by Rahnama (1390). The sensor consisted of three prismatic tips at equal distance of 10 cm, which were attached to three load cells by equaled-length push rods; the tips located at 10 and 30 cm deep were also equipped with microphones for recording the sound produced during the movement of the sensor through soil. It has been reported that the prismatic tip working at 10 cm depth, works through the disturbed soil produced by the sensor shank and the measurement with this tip did not have any significant relationship with the mechanical resistance measured by thr vertically-operated cone penetrometer for this layer (0-10 cm). Therefore, it was hypothesized that by increasing the length of the push rod for the tip working at this depth, the tip will penetrate through the undisturbed soil ahead of the disturbed soil. Also, in order to study whether the characteristics of the recorded sound is only a function of failure type that was claimed previously or depends on the failure type as well as the working depth of the tip, two more microphones were used and one was located on the sensor shank at 6 cm depth from the soil surface and the second one was inserted into the second tip located at 20 cm depth. The improved sensor was field tested at four water contents (0.5 PL, 0.7 PL, 0.9 PL and 1.1 PL (PL: Plastic limit)) for studying the sensor performance with long and short rods for the first tip located at 10 cm deep. In the second field test, the sensor was evaluated, using only the long push rod, at the same four water contents and 3 degrees of soil compaction created by tractor traffic (no pass, one pass and three pass). The results showed that there was a significant relationship between horizontal resistance index (HRI) and the mechanical resistance measured by a cone penetrometer. This significant relationship is probably due to the tip movement through the undisturbed soil ahead of rupture region created by the sensor shank. Both values of HRI and cone index (CI) significantly increased by a decrease in soil water content. In field evaluation of sensor at different level of soil water content and compaction, using long push rod for the 10-cm deep prismatic tip, significant relationships between HRI and CI for each depth were obtained. The analysis of sound signals recorded at three depths of 10, 20 and 30 cm showed that regardless of soil water content, increase in measuring depth, the power spectral density of the recorded sound was increased. Due to the failure mode at three measuring depths,this could be due to an increase in surcharge in soil with increasing depth. A significant relationship (R 2 = 0.74) between power spectral density and HRI at three depth was found. As a result, by using this developed sensor one can record the continuous variation in HRI in whole soil profile which has a significant relationship with CI. Key words: Horizontal resistance index; Cone index; Acoustic penetrometer