[Objective] The spatiotemporal variation and distribution of cultivated land quality at Minhou County, Fujian Province over many years were analyzed in order to provide new ideas and new help for the study of spatiotemporal pattern evolution of cultivated land quality in complex geomorphic areas in the southeast hilly region. [Methods] The spatiotemporal statistical scanning method was introduced into multi-time series cultivated land quality monitoring, and combined with the traditional spatial analysis method to study the continuous change trend and spatial distribution of county scale cultivated land quality. Furthermore, the experimental results generated by the introduction of the spatiotemporal scanning model were compared with the results of a local autocorrelation model in the traditional spatial autocorrelation analysis method. [Results] ① The spatiotemporal scanning model was superior to the local autocorrelation model for mining high-value clustering features. ② The change amplitude and spatial distribution breadth of cultivated land quality grades at Minhou County from 2018 to 2019 were greater than in 2019—2020. ③ From 2018 to 2020, the overall cultivated land quality at Minhou County was stable from year by year, and showed an upward trend. ④ According to the variation of the spatial and temporal patterns of cultivated land quality at Minhou County, cultivated land was divided into five areas: high quality improvement area, low quality improvement area, primary key improvement area, secondary key improvement area, and moderate improvement area. [Conclusion] Replenishing nutrients, increasing water conservation, and balancing soil pH value are important practices to implement to prevent and reverse the deterioration of cultivated land quality and to maintain high quality and high yield of farmland. In addition, establishing cultivated land protection areas to prevent non-agricultural uses of cultivated land is also conducive to stabilizing and improving cultivated land quality.