Background: Pleuropulmonary blastoma (PPB) is a rare lung malignancy in children derived from mesenchyme. Single-cell RNA sequencing is used to explore the heterogeneity of tumors. Here, we used this technology to classify PPB subpopulations and predict their differentiation trajectory. Procedure: This study included 10 007 single cells from a girl with PPB. After choosing malignant tumor cells with an inferring copy number variation, we used non-negative matrix factorization and Seurat analysis to cluster the cells and divided the subgroups by their differentially expressed genes and Gene Ontology enrichment analysis. Additionally, pseudotime trajectory analysis of PPB was conducted with Monocle. Results: Tumor cells were divided into two major categories including muscle (DEShiTNNT1hiTNNI1hi) and cartilage (TWIST1hiHTRA1hiBMP4hi). In muscle lineage, satellite myogenic cells (PAX7hiMYF5hiMSChi), which were subdivided into more primitive FABP7hi one and later ITM2Ahi one, were the origin of muscle in PPB and gradually differentiated into myocytes (MYOGhiTTNhiMYL4hi). And in the cartilage lineage, SOX9hiPAX1hiPAX9hi prechondrocytes generated MGPhiOGNhi chondrocytes and DKK2hiTNMDhi chondrocytes. Conclusions: Our results demonstrated unrecognized heterogeneity in PPB at the single cell level and defined its muscle and cartilage subpopulations together with their trajectories, which would provide some basis for the molecular mechanism and treatment of PPB.