Objective To develop and validate a high-sensitivity detection method for Salmonella in raw vegetables using a large sample size （500 g）， and to characterize the isolates from retail raw vegetables in Beijing.Methods The effect of different sample sizes on Salmonella recovery from artificially contaminated lettuce and coriander was analyzed， and the isolation efficiency from retail samples in Beijing was compared between 25 g and 500 g sample sizes. The antimicrobial susceptibility of Salmonella isolates to 16 antibiotics was tested， and serotypes and resistance genes were analyzed by whole genome sequencing （WGS）.Results Increasing the sample size improved detection sensitivity （0.002 6 CFU/g） in raw vegetables， and achieving a limit of detection of 10⁰ CFU/500 g. Testing of retail samples in Beijing demonstrated that the method using 500 g sample significantly increased the Salmonella detection rate （8.1%） compared to the standard method using 25 g sample （0.5%） （P<0.05）. The contamination distribution of Salmonella in raw vegetables exhibited significant regional and retail-type variations （P<0.05）， higher in suburban areas （11.2%） than in urban areas （3.5%）， and higher in grocery （11.4%） and open markets （8.8%） than in supermarkets （6.1%） and greengrocers （0%）. Among the 18 Salmonella isolates， 12 serotypes were identified. S. Alachua， the major serotype， represents the first food-borne isolate of this serotype reported in China. Overall， 33.3% of isolates were resistant to at least one antibiotic， and 22.2% were multidrug-resistant. WGS identified 20 resistance genes， including β-lactamase and quinolone resistance genes.Conclusion Increasing the sample size to 500 g overcomes the challenge of missing low-level Salmonella contamination in raw vegetables， significantly enhancing the detection rate and providing the high-sensitivity screening strategy. Salmonella isolates from raw vegetables in Beijing exhibited serotype diversity and multidrug resistance.