Zebrafish exhibit a robust ability to regenerate their hearts following injury, and the immune system plays a key role in this process. We previously showed that delaying macrophage recruitment by clodronate liposome (–1d_CL, macrophage-delayed model) impairs neutrophil resolution and heart regeneration, even when the infiltrating macrophage number was restored within the first week post injury (Lai et al., 2017). It is thus intriguing to learn the regenerative macrophage property by comparing these late macrophages vs. control macrophages during cardiac repair. Here, we further investigate the mechanistic insights of heart regeneration by comparing the non-regenerative macrophage-delayed model with regenerative controls. Temporal RNAseq analyses revealed that –1d_CL treatment led to disrupted inflammatory resolution, reactive oxygen species homeostasis, and energy metabolism during cardiac repair. Comparative single-cell RNAseq profiling of inflammatory cells from regenerative vs. non-regenerative hearts further identified heterogeneous macrophages and neutrophils, showing alternative activation and cellular crosstalk leading to neutrophil retention and chronic inflammation. Among macrophages, two residential subpopulations (hbaa+ Mac and timp4.3+ Mac 3) were enriched only in regenerative hearts and barely recovered after +1d_CL treatment. To deplete the resident macrophage without delaying the circulating macrophage recruitment, we established the resident macrophage-deficient model by administrating CL earlier at 8 d (–8d_CL) before cryoinjury. Strikingly, resident macrophage-deficient zebrafish still exhibited defects in revascularization, cardiomyocyte survival, debris clearance, and extracellular matrix remodeling/scar resolution without functional compensation from the circulating/monocyte-derived macrophages. Our results characterized the diverse function and interaction between inflammatory cells and identified unique resident macrophages prerequisite for zebrafish heart regeneration.