Abstract: This study was dedicated to developing a visualization technique based on loop-mediated isothermal amplification (LAMP) with novel molecular targets, aiming to simultaneously, rapidly, and sensitively detect Flavobacterium spp. in fresh fish meat. The whole-genome sequencing and pan-genome analysis techniques were employed to systematically mine the candidate target of the Flavobacterium genus. Five Flavobacterium strains and forty-four non-Flavobacterium strains were prepared for specificity verification of candidate target. LAMP primers were designed based on the target sequence and then the amplification system was optimized. The specificity and sensitivity of novel LAMP assay and LAMP based lateral flow dipstick (LFD) visualization method were evaluated. Finally, the anti-interference capability and the accuracy in detecting actual samples were determined. These results showed that the menB target mined in this study exhibited good specificity, and the newly established LAMP method had no cross-reactivity against 12 non-target bacteria trains in optimal conditions. The detection limits for Flavobacterium in pure cultures and mixed spiked fish meat samples were 3.3×10² CFU/mL and 3.3×10³ CFU/g, respectively. Meanwhile, the detection sensitivity and specificity of LAMP-based visualization technique in pure cultures and mixed spiked fish meat samples were comparable to those of the traditional LAMP method. Moreover, it maintained excellent detection performance even under the interference of high concentrations of contaminant bacteria and spoiled fish sample. The detection results of the LAMP-LFD visualization methods for 24 actual samples were consistent with those of the traditional culture method. The novel LAMP visualization technology established in this study provides an efficient and reliable detection method for the prevention and control of Flavobacterium contamination in food, which has important application value and promotion prospects.