Recently, the construction of renewable energy generation systems in uninhabited areas such as deserts and oceans has increased significantly, along with the low-carbon transition of the global energy system. However, the traditional planning methods that rely on information from nearby meteorological stations often yield suboptimal planning schemes due to the vastness of the uninhabited areas and significant meteorological spatiotemporal differences. Therefore, an information value-driven comprehensive benefit evaluation method is proposed in this article to efficiently and accurately evaluate suitable locations for renewable power generation system. The information matrix of geo-meteorological coupling is constructed through the numerical weather prediction correction method of Fourier ForeCasting Neural Network. To replace the traditional time-consuming planning method based on traversal optimization, a comprehensive evaluation method that fully considers the value of information itself and its multi-dimensional values such as economy, safety, and low-carbon for the planning scheme is proposed in this paper. Case study based on the Tengger Desert in China is carried out to verify the efficiency of the proposed method.