Background Indocyanine green fluorescent lymphography helps visualize the lymphatic drainage pattern in gastric cancer; however, it is unknown whether fluorescent lymphography visualizes all metastatic lymph nodes. This study aimed to evaluate the sensitivity of fluorescent lymphography to detect metastatic lymph node stations and lymph nodes and the risk of false-negative findings. Methods Patients with clinical T1–4a gastric cancer were included. Indocyanine green was peritumorally injected the day prior to surgery by endoscopy. Gastrectomy with systematic D1+ or D2 lymphadenectomy was performed. Stations and lymph nodes were retrieved at the back-table using near-infrared imaging and classified as “fluorescent” or “non-fluorescent” and later matched with histopathological findings. Results Among 592 patients who underwent minimally invasive gastrectomy from September 2013 until December 2016, lymph node metastases were present in 150. The sensitivity of fluorescent lymphography in detecting all metastatic lymph node stations was 95.3% (143/150 patients), with a false-negative rate of 4.7% (7/150 patients) and the sensitivity in detecting all metastatic lymph nodes was 81.3% (122/150 patients). The negative predictive value was 99.3% for non-fluorescent stations and 99.2% for non-fluorescent LNs. For detecting all metastatic LN stations, subgroup analysis revealed 100% sensitivity for pT1a, 96.8% for pT1b, 100% for pT2, 91.3% for pT3, and 93.6% for pT4a tumors. Conclusions Fluorescent lymphography-guided lymphadenectomy can be a useful method for radical lymphadenectomy by facilitating the complete dissection of all potentially positive LN stations. Fluorescent lymphography-guided lymphadenectomy appears to be a reasonable alternative to conventional systematic lymphadenectomy for gastric cancer.