The invention disclose a protein fingerprint atlas-spectrum model of acidovorax avenae subsp. citrulli, the model is drawn according to the mass-to-charge ratios of acidovorax avenae subsp. citrulli protein and wave crest strength coefficients thereof, wherein the mass-to-charge ratios comprises: 3171.4, 3735.0, 4963.5, 5139.0, 5866.4, 6554.1, 7466.1, 7948.0, 8384.5, 9304.2, 9504.3, 9920.0, 10272.8, 10840.7, 12530.5, and 12896.5. The invention also discloses a matrix-assisted laser desorption/ionization flying time mass spectrum detection method based on the model. The method can qualitatively detect whether a sample contains acidovorax avenae subsp. citrulli. The model has a function of specific recognition, is capable of being applied to commercial detection of acidovorax avenae subsp. citrulli, and is especially suitable for rapid and high throughout detection for departments such as port inspection and quarantine.