Taking Fujian Province as the research object, based on the data from 37 national air monitoring stations and land use/cover data in Fujian Province, the effects of land use/cover on the annual and seasonal variations of SO2, NO2, O3 and CO pollution concentrations were analyzed. Then, buffer zones of different radii of the national monitoring stations were established, the landscape pattern indexes were calculated, and the effects of the landscape pattern of land use/cover on SO2, NO2, O3 and CO concentrations at different scales were discussed. The results showed that: Land use/cover had a significant effect on the change of atmospheric pollutant concentration. SO2, NO2 and CO concentrations all showed the highest values in construction land, and O3 concentrations showed the highest values in cultivated land. The seasonal variation of the concentration of air pollutants under different land uses/covers was different. The concentration of SO2 in cultivated land was low in spring and winter, high in summer and autumn, and others were high in spring and winter, low in summer and autumn. The concentration of NO2 was high in spring and winter, but low in summer and autumn. The concentration of O3 in cultivated land and forest land was high in spring and autumn, and low in summer and winter, while the concentration of O3 in construction land and grassland decreased from spring to winter successively. The CO concentration was low in summer and autumn, but high in spring and winter. The effects of different landscape pattern indexes on the concentration of air pollutants were different with the scale effect. Among them, the patch density (PD) in the grassland within the radius of 3 000 m in spring and winter and 4 000 m in summer and autumn had the most significant and negative correlation on SO2 concentration, indicating that the higher the density of grassland patches was, the lower the SO2 concentration would be. There was a positive correlation between the number of patches (NP) and NO2 concentration in the 4 000 m radius of forest land in spring and winter, indicating that the more broken forest land was, the higher the NO2 concentration would be. NP of construction land within 3 000 m radius was negatively correlated with NO2 concentration, indicating that the more broken the construction land was, the lower the NO2 concentration would be. There was a positive correlation between the concentration of O3 and the proportion of landscape (PLAND) of cultivated land in the radius of 5 000 m, indicating that the greater the PLAND of cultivated land was, the higher the concentration of O3 would be. There was a positive correlation between the concentration of O3 and the PLAND of forest land in the radius of 1 000 m, indicating that the increase of PLAND of forest land had certain influence on the increase of O3 concentration. Except for autumn, the PLAND of forest land was related to the CO concentration in the radius of 1 000 m, showing that the greater the PLAND of forest land was, the lower the concentration of CO would be. At the same time, through the analysis of different scales of landscape indexes, it was found that the best study scale of SO2 was 3 000 m, and that for NO2 and O3 was 4 000 and 5 000 m, respectively, while a better study scale for CO was 3 000 m.