In sandy, gravelly and desert regions, the electricity-heat-hydrogen multi-energy coupling systems have been proven effective in promoting renewable energy accommodation of renewable energy bases. To explore efficient and economical production and utilization pathways for electricity, heat, and hydrogen, this paper proposes a day-ahead distributionally robust optimal (DRO) scheduling strategy for such multi-energy coupling systems tailored to sandy, gravelly, and desert regions. Firstly, an electricity-heat-hydrogen supply architecture for the multi-energy coupling system without conventional power supply support is proposed. Considering the heat exchange cycle and air cycle, refined models are formulated for multi-energy production and utilization links as well as the key equipment of the electricity, solar thermal, and hydrogen subsystems. Besides, it deeply couples battery energy storage (BES), thermal energy storage (TES), and hydrogen storage (HST) for multi-energy scheduling. To deal with the uncertainty of renewable generation, a data-driven day-ahead DRO scheduling model is established where the Wasserstein distance is adopted to construct an uncertainty fuzzy set for renewable energy prediction errors. Moreover, the DRO model is transformed into a deterministic model with nonlinear constraints via strong duality theory, and the adaptive McCormick envelope is employed to handle the nonlinear constraints. Finally, comparing and analyzing different operating scenarios, the proposed system architecture has absorbed 98.6% of renewable energy, improved the energy supply stability by 4%, and the DRO scheduling model has well balanced robustness and economy. Notably, the proposed adaptive McCormick envelope method strictly controls the nonlinear approximation error within 1.62% while accelerating the computational speed by 3.6 times compared to traditional standard methods.