Spatially resolved transcriptomic technologies have emerged as pivotal tools for elucidating cellular heterogeneity and molecular regulation within the complex tissue microenvironment, but are constrained by insufficient gene recovery or an inability to achieve intact single-cell resolution. By integrating single-cell RNA-seq and spatial transcriptomics, we develop a mathematical method of single-cell resolved spatio-temporal (SCST) mapping that comprises tiered algorithms for constructing the spatial molecular atlas of the biospecimen at single-cell resolution across a timeline of development. Each step of SCST method can be applied independently in spatial omics study. The embedded spatial-smoothing algorithm in SCST significantly enhances the spatial mapping accuracy of single cells, thereby improving the fidelity of the annotation of cell identity to the equivalent in vivo cell type. Through 3D mathematical modelling, SCST facilitates the spatial reconstruction of single-cell molecular atlas and the delineation of cellular heterogeneity. When integrated with temporal data, SCST can also delineate the spatio-temporal lineage trajectory at single-cell resolution in a developing biological entity.