<p>A small fraction of high-emitting vehicles make disproportionally large contributions to total fleet emissions. Therefore identifying high emitters under real driving conditions is crucial. In this study, two portable sensor platforms for high-emitter identification were used for online roadside measurements of vehicle-emitted NO, particle number (PN), and CO₂ concentrations in Tangshan and Chengdu, respectively. The measured mean concentrations of vehicle-emitted NO, PN, and CO₂ in Tangshan and Chengdu were 27.7–32.9 ppb, 5.4 × 10³–8.2 × 10³ #/cm³, and 7.3–8.2 ppm, respectively. Based on more than one month of second-by-second measured pollutant concentrations and passed vehicle information, a scheme was developed to identify high emitters. Among the 217000 and 43000 vehicles that passed the roadside sensor platforms at Tangshan and Chengdu, approximately 60% and 73% of vehicle exhaust plumes were successfully detected using the sensor platform. The NO and PN emission factors (EFs) tended to have log-normal distributions with the median values of 14.3 g/kg-fuel and 1.3 × 10¹⁵ #/kg-fuel, respectively. In general, the percentages of high-emitters identified at the Tangshan and Chengdu sites were 8.7% and 12.2% of the total identified vehicles, respectively. Among these high-emitters, 122 vehicles were randomly inspected on-site with the assistance of traffic officers, and the rate of correct identification was approximately 95%, which demonstrates that our methodology performs well in identifying real-world high-emitters. Overall, its low cost, good mobility, strong adaptability, and high correct identification rate make this roadside sensor platform a promising approach for real-world high-emitter identification.</p>