The use of spatio-temporal statistical combination method to explore the land use transition in urban suburbs provides an essential resource for promoting urban-rural integration and management of further urbanization. A combination of Space Time Cube and Geodetector methods was used to quantitatively identify the dominant morphology of land use change, and explore the recessive morphology and driving mechanism in Baiyun District, a suburb of Guangzhou from 1980 to 2020. Between these years, the land use changes were dominated by the expansion of built-up land and the shrinkage of cropland. The expansion took place in southern (1990-2000) to central (2000-2005) areas of the jurisdiction, followed by large-scale decentralization (2005-2010). Land use activities showed a downward trend. A continuous cold spot was mainly distributed in the southeast and remote areas near the urban center, whereas a continuous hot spot was spatially characterized by a multinuclear pattern, mainly concentrated in the center and south. Socioeconomic factors significantly affected the spatial pattern of land use activities. The trend turning in land use morphology of land use change can be divided into four stages: land primitivation from 1980 to 1990, land fragmentation from 1990 to 2000, land overflow from 2000 to 2010, and land intensification from 2010 to the present. Baiyun District is at the forefront of China’s development and has typical “suburban” characteristics, which can provide a reference for other suburbs in China to develop in stages. Today, urban suburbs in China face the problems of limited land resources and mixed land use space. Taking advantage of the new national strategic environment could effectively solve these land use issues in urban suburbs and achieve sustainable development. New trends in land use morphology research, also considering land use effects and land use change mechanisms in an integrated manner, are conducive to further deepening the understanding of land use transitions. Moreover, spatiotemporal statistical methods have innovated the field of mining the spatiotemporal characteristics of land use. Based on a scientific method for spatiotemporal analysis, we were able to reconstruct the three-dimensional space of land and improve the methods applied for research into land use change.