Based on data from 286 cities, this paper analyzes the realistic effect of natural resource supervision on construction land use efficiency via the difference-in-differences (DID) method and its potential effect via stochastic frontier analysis (SFA). First, in the real economy, natural resource supervision can significantly reduce construction land use efficiency in cities with a natural resource supervision agency (NRSA), but the inhibitory effect gradually weakens. Natural resource supervision can promote the market development of tender, auction and quotation (TAQ) and thereby have an indirect impact by strengthening illegal land management. Second, natural resource supervision can significantly improve construction land use efficiency in cities with an NRSA when the degree of vertical financial asymmetry is high, but it has the opposite effect when the fiscal revenue decentralization level, regional competition level, economic development level and so on are high. Third, in terms of the potential effect, construction land use efficiency in cities without an NRSA is decreasing annually. Natural resource supervision can greatly promote construction land use efficiency in cities with an NRSA and allow these cities to catch up with cities without an NRSA. From the logical starting point of the resident effect of natural resource supervision, this paper focuses on the ultimate goal of high-quality economic development. This study is of great value for reforming the natural resource supervision system, standardizing the development of the land market and perfecting the incentive mechanism for local governments.