This paper describes a study on the design, fabrication and testing of a prototype digging device for sweet potato tubers in bris soil. The soil texture was sandy soil (fine sand 94.53%), with mean moisture content of 9.16% and mean bulk density of 1.44 g-cm-3. The soil was prepared in a soil bin. Three types of soil digging tools were designed and fabricated to determine the optimum draft force. These were Flat or plane, V-shaped and Hoe type blades. Plane and V-shaped blades were 30 cm long, and 13 cm wide, while the Hoe type had three rods, 25 mm in diameter, 30 cm long and 6.5 cm wide with sharp cutting edge. The digging tools were tested in a soil bin filled with bris soil to determine the optimum draft force and area of soil disturbance. The results were analysed using statistical analysis of variance (ANOVA). Comparison between all blade types and blade depths to measured draft force and the area of soil disturbed showed that the highest draft of 0.54 kN-m-2 was caused by a flat or plane blade at the optimum depth of 20 cm when the area of soil disturbed was 0.180 m2 . The V-shaped blade had the mean draft of 0.51 kN-m-2, with area of soil disturbance of 0.185 m2 . Thebest choice was V-shaped blade with a rake angle of 30o at 20 cm. depth. The selected blade was fixed onto the sweet potato harvester and tested on bris soil planted with sweet potato of Telong and VitAto varieties. The harvesting efficiency of the machine in bris soil was 93.64% and 90.49% for Telong (Plot A) and VitAto (Plot B) varieties, respectively. The average ground speed and turning time during operation for plots A and B was 0.56 km-hr-1 and 102.7 s and 0.99 km-hr-1 and 81.22 s, respectively. The harvesting efficiencies for both plots showed no significant difference. The total productive time (harvesting time) and unproductive time (turning time) in plot A, at a tractor speed of 0.56 km.hr-1, was 14.8 hours for harvesting a hectare of sweet potato ( 0.068 ha.hr-1). In plot B, the total time for harvesting a hectare of sweet potato was 8.35 hours (0.12 ha.hr-1) at a tractor speed of 0.99 km.hr-1. The average harvesting time for both plots was 11.47 hr.ha-1. The average field work rate was 0.087 ha.hr-1 or 34 man-hr.ha-1 compared to manual harvesting of 150 man-hrs.ha-1.