Physiological and biomechanical strains were investigated in women using four different handle designs for hoe. Electromyography (EMG), energy expenditure, and task workload were measured while using four different hoes: a conventional hoe with 1) the normal (N) handle, 2) the second (S) handle, 3) the T-grip with arm supporter (TA) handle, and 4) the second with T-grip and arm supporter (STA) handle. Each handle strained different regions of the body. The S handle was significantly better for minimizing the erector spinae muscle EMG. The TA handle required less effort on the flexor digitorum profundus (FDP), but more effort on the posterior deltoid. From the STA handle, no interaction was found between the S and TA handle. The results from energy expenditure analysis indicated that the workload in hoeing involves mainly a pushing and pulling motion of the shoulder (r = .923). In comparison with the N handle, the additional handles caused a significantly greater workload, but there was no significant difference between those three handles.