We assessed the influence of different directions of suture strands using the double Tsuge repair on the resistance to gap formation and mode of failure. In total, 25 porcine flexor digitorum profundus tendons were split into two groups. One group was repaired using a conventional double Tsuge suture created by two bands of looped suture placed longitudinally parallel (parallel method), and the other group was repaired using a new technique comprising two bands of looped suture placed in a crossed fashion in the anterior and posterior half of the tendon (cruciate method). The repaired tendons were subjected to linear, non-cyclic load-to-failure tensile testing. The cruciate method had a significantly higher mean load at a 2-mm gap tensile load (29.7 N [SD, 8.3]) than the parallel method (21.6 N [SD, 4.9]) and failed significantly more often due to suture pull-out. The direction of a core suture and its location within the tendon affect both gap resistance and the mode of failure of repair when using the double Tsuge suture technique, with a cruciate configuration achieving a greater gap resistance than a parallel one.