Single-cell RNA sequencing (scRNA-seq) was used to analyze the developing mouse molars, in order to construct a spatiotemporal development atlas of pulp cells, and further to reveal the developmental process and regulatory mechanism of tooth development. Ten mandibular first molars from C57BL/6 mice in postnatal day (PN) 0 and 3 were respectively dissected and digested to obtain single-cell suspensions. scRNA-seq was performed on 10× Genomics platform. PN 7 mouse molar scRNA-seq data were obtained from our previous study. PN 0, 3, and 7 scRNA-seq data were integrated for following analysis. The initial quality control, mapping and single cell expression matrix construction were performed by Cell Ranger. Quality control, standardization, dimensional reduction and cluster analysis were performed by using Seurat. Monocle was used to generate the pseudotime trajectory. Scillus was used to perform gene ontology analysis. In order to detect the spatiotemporal change of different population of pulp cells, the marker genes of each cluster were demonstrated by RNAscope in situ hybridization. There were twenty-six cell clusters within mouse molars, which were identified as eight different cell types, including dental pulp cells, dental follicle cells, epithelial cells, immune cells, endothelial cells, perivascular cells, glial cells and erythrocytes. We further re-clustered and analyzed dental pulp cells. Cluster 0 were mature pulp cells, which located at the upper portion of crown. The main functions of cluster 0 were osteogenesis and extracellular structure organization. Cluster 1 were apical papilla cells, which located at the apical part of roots, whose main functions were extracellular structure organization and organ development. Cluster 2 were cycling cells, which were actively proliferated, resided in the lower portion of the crown. Cluster 3 and 4 were preodontoblasts and odontoblasts, respectively. Their functions were closely related to biomineralization. The proportion of mature pulp cells increased with the development process, while the proportion of cycling cells and odontoblast lineage decreased. According to the expression pattern of marker genes of each cluster, we constructed a cell atlas of dental pulp. Pseudotime trajectory analysis found there were two development trajectories within dental pulp. They both started from SPARC related modular calcium binding 2 (Smoc2)+ dental papilla cells, then went through DNA topoisomerase Ⅱ alpha (Top2a)+ cycling cells, and finally divided into coxsackie virus and adenovirus receptor (Cxadr)+ mature pulp cells or dentin sialophosphoprotein (Dspp)+ odontoblasts two lineages. scRNA-seq could fully discover the intercellular heterogeneity of cells on transcriptome level, which provides a powerful tool to study the process and regulatory mechanism of organ development.