Targeted imaging of cancer lymphatic metastasis remains challenging due to its highly heterogeneous molecular and phenotypic diversity. Herein, we introduced triple-targeted protein nanoprobes capable of specifically binding to three targets for imaging cancer lymphatic metastasis, through a data-driven design approach combined with a synthetic biology-based assembly strategy. Specifically, to address the diversity of metastatic lymph nodes (LNs), a combination of three targets, including C-X-C motif chemokine receptor 4 (CXCR4), transferrin receptor protein 1 (TfR1) and vascular endothelial growth factor receptor 3 (VEGFR3) was identified, leveraging machine leaning-based bioinformatics analysis and examination of lymphatic metastatic tissues from patients with gastric cancer. Using this identified target combination, ferritin nanocage-based nanoprobes capable of specifically binding to all three targets were designed through the self-assembly of genetically engineered ferritin subunits using a synthetic biology approach. Multiplexed imaging of heterogeneous metastatic LNs using these nanoprobes resulted in approximately 80% binding with heterogeneous tumor cells, which was 60% higher than that of single-targeted nanoprobes in a polyclonal lymphatic metastasis animal model. In 19 freshly resected human gastric specimens, the signal from the triple-targeted nanoprobes significantly differentiated metastatic LNs from benign LNs, in high accordance with pathological examination. This study not only provides an effective nanoprobe for imaging highly heterogeneous lymphatic metastasis but also proposes a potential strategy for guiding the design of targeted nanomedicines for cancer lymphatic metastasis. This article is protected by copyright. All rights reserved.