Aiming to address the issues of high resistance and energy consumption during rotary tillage in tea gardens of southern hilly and mountainous areas, a drag-reducing design for a ditching blade was developed based on the bionic physical characteristics of mole forefeet. Reverse engineering technology was employed to extract the tip and lateral contour features of the middle three toes (the 2nd, 3rd, and 4th toes) of the mole forefoot. Gaussian equation and the fifth-order polynomial equation were used for curve fitting (R2≥0.9986, SSE≤0.0085), followed by wedge simplification and tooth arrangement optimization. Based on the fitting results, three types of bionic toothed rotary blades (types A, B, and C) were designed. A tool-soil interaction simulation model was established in the EDEM 2023 discrete element software, and the average torque and soil fragmentation rate of the conventional blade and the three bionic rotary tiller blades under different parameter conditions were obtained. Multi-condition performance verification was further carried out via bench tests and field experiments, in which the forward speed was 0.5m/s, the rotary tillage depth was 120mm, and the blade shaft rotational speeds were set to be 150r/min, 200r/min, and 250r/min, respectively. Simulation results showed that the average torques of the three bionic rotary tiller were reduced by 6.40% , 3.73% and 5.08% compared with that of the conventional rotary tiller blade, and the soil breaking rates were increased by 7.09, 8.82 and 4.68 percentage points, respectively. The optimal drag reduction performance was obtained for type A bionic rotary tiller blade, and the best soil crushing capacity was achieved by type B bionic rotary tiller blade. Bench test results indicated that the average torques of the three bionic rotary tiller blades were decreased by 7.25% , 4.01% and 5.49% in comparison with the traditional cutter. Field test results demonstrated that the average soil breaking rates of bionic rotary tiller blades A, B and C were measured as 91.68% , 92.84% and 89.86% , which were improved by 5.39, 6.55 and 3.57 percentage points relative to that of the traditional cutter.