Vehicle emission has been regarded to the major source of air pollution in urban area. Especially, the emissions originating from container diesel trucks (CDTs) constitute a significant portion of road transportation emissions in port city. However, the instantaneous measurement of CDTs emission in real road situation remains limited. In this study, we employed a portable emission measurement system (PEMS) to measure the real-world emissions of CDTs that meet the China IV and China V standards on highway and port roads. Based on the data, the impacts of vehicle speed, acceleration, and vehicle specific power (VSP) on CO2 and NOx emission rates are investigated. Additionally, random forests is employed to explore the interplay of driving, engine, and external factors on vehicle emission rates, as well as to predict these rates. Finally, the emission factors under diverse driving conditions are examined. The results indicate that elevated CO2 and NOx emission rates are primarily observed during instances of low and medium-speed acceleration on highway. On port road, the distribution of high CO2 and NOx emission rates are relatively scattered. The relationship between the emission rates of CO2 and NOx and VSP follows a third-order polynomial function. Engine power, engine speed, and VSP exhibit a strong correlation with CO2 and NOx emission rates. Remarkably, the CO2 and NOx emission factors on highway for China V CDT have decreased by 58.04% and 55.13%, respectively, compared to China IV CDT. Display omitted •High emission of CO2 and NOx appeared during acceleration at low and medium speeds.•CO2 and NOx emissions strongly link to VSP in a third-degree polynomial.•Driving factors explain 63% of CO2 emissions in China V container diesel trucks.•The China V container diesel truck emits 46.6% less CO2 than the China IV truck.