Paraspinal muscle fatigability during various trunk extension tests has been widely investigated by electromyography (EMG), and its task-dependency is established recently. Hip extensor muscle fatigability during the Sorensen test has been reported. The aim of the present experiments was to evaluate the task-dependency of back and hip extensor muscle fatigue during two variants of the Sorensen test. We hypothesized that the rate of muscular fatigue of the hip and back extensor muscles varies according to the test position. Twenty healthy young males with no history of low back pain volunteered to participate in this cross-sectional study. They were asked to perform two body weight-dependent isometric back extension tests (S1 = Sorensen test; S2 = modified Sorensen on a 45 degrees Roman chair). Surface EMG activity of the paraspinal muscles (T10 and L5 levels) and hip extensor muscles (gluteus maximus; biceps femoris) was recorded, and muscular fatigue was assessed through power spectral analysis of the EMG data by calculating the rate of median power frequency change. We observed hip extensor muscle fatigue simultaneously with paraspinal muscle fatigue during both Sorensen variants. However, only L5 level EMG fatigue indices showed a task-dependency effect between S1 and S2. Hip extensor muscles appear to contribute to load sharing of the upper body mass during both Sorensen variants, but to a different extent because L5 level fatigue differs between the Sorensen variants. Our findings suggest that task-dependency has to be considered when EMG variables are compared between two types of lumbar muscle-fatiguing tasks.