Treating metastatic tumors at micrometastatic stage is a potential strategy to slow cancer progression. This study aims to establish an apoptosis sensor-based platform for rapid, effective and non-invasive identification of drugs that inhibit proliferation of micrometastatic cancer cells. We stably transfected a plasmid encoding the fluorescence resonance energy transfer (FRET)-based caspase-3 sensor into high-metastatic melanoma B16F10 cells. The resulting B16F10-C3 cells were applied for screening of anti-proliferative and pro-apoptotic drugs in two-dimensional (2D) monolayer, three-dimensional (3D) spheroids, and zebrafish xenotransplantation tumors. All studies were conducted in 96-well plates in a high-throughput manner. 14 compounds including 6 chemotherapy drugs and 8 kinase inhibitors were tested. 13 compounds failed the tests due to different reasons: drug resistance, low efficacy, poor pharmacokinetic profile or high toxicity to zebrafish. Only one compound, pan-PI3K inhibitor LY294002, passed all tests. It not only inhibited the proliferation of B16F10-C3 cells in 2D and 3D cultures, but also significantly reduced the growth of xenograft tumour in zebrafish. Our study suggests that PI3K/AKT pathway is a potential therapeutic target for the reactivation of tumour dormancy and proliferation of micrometastases. Moreover, this integrated approach is effective for rapid identification of systemic anti-metastases drugs.