Abstract: Agricultural machinery and equipment are indispensable to realize resource utilization for high efficiency in sustainable production. It is also required for the efficient, intelligent, and environmentally friendly agricultural power machinery, due to the greenhouse gas emissions and the drastic reduction in the extraction of non-renewable resources. Fuel cell distributed drive electric tractors (FCDET) can provide a new approach to developing green agricultural machinery. However, the great challenge has posed on the traction efficiency, short range, and high hydrogen consumption. A reasonable and effective power allocation can be expected to reduce energy consumption for the high efficiency of the fuel cell system. In this study, an allocation strategy of plowing drive power was proposed for FCDET using adaptive multi-resolution analysis (AMRA). The effective decoupling between the various energy sources was also realized to reduce the frequent start-stop and the large power fluctuations of the hydrogen fuel cell. Firstly, two models were established for the fuel cell system and the total efficiency solver. The first-time reconstruction was to obtain the subsequence that effectively responded to the oscillation characteristics of the power signal using the tunable Q-factor wavelet transform (TQWT). The second reconstruction was to decompose the low-frequency subseries into several discrete sub-signals with special sparse properties using variational mode decomposition (VMD). Then, the sparrow search algorithm (SSA) was used to obtain the optimal combination of modal decomposition layers and quadratic penalty factors for VMD in real time. Finally, the second-time decomposed subsequence and sub-signals were reconstructed, according to the frequency characteristics. The reconstructed power signal was redistributed among the various energy sources. A test was carried out to verify the power allocation. The power demand information of the drive motor was acquired for FCDET plowing conditions at the China YTO Mengjin test base, taking the ET504-H prototype as the object. The plowing condition included a 0-8.7 s starting stage and an 8.7-18.0 s stable plowing stage. The demand power signal shared a large rate of change in the starting stage and a large high-frequency characteristic in the stable plowing stage. In addition, another experiment was carried out on the transmission bench in the New Energy Key Laboratory of Henan Province, in order to test the cooperative operation of permanent magnet synchronous motors (PMSM) under three power allocations. The results showed that the FCDET drive power allocation using AMRA effectively improved the energy utilization of the fuel cell system and the economy of the whole vehicle in plowing condition. The fuel cell system efficiency was improved by 8.30% and 1.82%, respectively, compared with the power-following and the first-time reconstruction. The equivalent hydrogen consumption was reduced by 35.60% and 11.86%, respectively. Meanwhile, the drive motor efficiency was improved by 2.06% and 1.27% on average; the energy consumption was reduced by 3.73% and 2.60%, respectively. This finding can provide a novel theoretical and technical approach for the development of the FCDET control system.