<p>Heart development involves dynamic signaling interactions between cells and their surrounding environment (niche). Single-cell mRNA sequencing (scRNA-seq) has been widely used to profile gene expression in individual cells, but it faces challenges in dissecting niche signals due to the need for cell dissociation. In contrast, spatial transcriptomics can preserve tissue structure and represents a potentially effective approach for this purpose. In this study, we used two spatial transcriptomics platforms, 10x Genomics Visium and Curio Slide-seq (Curio Seeker), to generate a spatial atlas of hearts at embryonic and neonatal stages. Using Visium data, we analyzed the spatial patterns of cell cycle phases, compact and trabecular myocardium signatures, and chamber-specific genes across developmental progression. We discovered that atrial cardiomyocytes exhibit a mature myocardium transcriptional signature. Additionally, we identified the spatial patterns of signaling activities at different stages. Using Slide-seq data, we identified cardiac conduction cells, including cardiac neurons, sinoatrial nodal cells, atrioventricular nodal cells, and Purkinje fiber cells, and further studied their niche signaling. Moreover, by combining lineage tracing and spatial transcriptomics, we identified four types of epicardial cell-derived cells (EPDCs) and analyzed their signaling interactions with niche cells. We then eliminated the EPDCs using a cell ablation system and observed reduced signaling in the ablated hearts through spatial transcriptomics analysis. In summary, we generated a spatial transcriptomic atlas for developing mouse hearts and identified niche signaling for cardiac conduction cells and EPDCs.</p>