Abstract: Until now, few works have reported comparison results between subjective and objective air quality forecast. To understand the performances of national subjective and objective air quality forecast, operational results in 2018 from National Meteorological Center were verified by threat score (TS) method. The results showed that TS score, false alarm ratio and missed alarm ratio of subjective (objective) forecast for mild and above pollution all over China in 2018 were 0.23-0.34 (0.24-0.26), 0.37-0.43 (0.39-0.41) and 0.58-0.72 (0.68-0.71), respectively. This illustrated that subjective forecast was better than objective one, but the ability of objective forecast was close to subjective forecast. Performances of subjective and objective forecast were both better in regions with heavier pollution (Jing-Jin-Ji, Fenwei Plain, Central China and Yangtze River Delta) than in lower pollution ones (Northwestern China, Southwestern China, Pearl River Delta and Northeastern China), and so did it in the heavily polluted winter than in the clean summer. With the extension of forecast time, TS scores of subjective forecast showed a downward trend, but scores of objective forecast did not change too much. In winter with heavy pollution, TS scores of 48 and 72 hours' objective forecast were higher than those of subjective forecast. Additionally, in 5 haze processes in 2018, TS scores of subjective (objective) forecast for mild, moderate and severer pollution all over China were 0.39-0.57 (0.43-0.46), 0.22-0.46 (0.25-0.30) and 0.10-0.34 (0.10-0.18), respectively. The situation of TS scores of subjective forecast higher than objective forecast accounted for 3 times in 5 processes. But performance of objective forecast was more stable with extension of forecast time, which illustrated that objective forecast could provide stable reference to forecasters in pollution process, but its ability in dealing with high pollution needs to be improved.