Increasing evidence has shown that exposure to air pollution is linked to adverse birth outcomes, but the results are not consistent. This study was performed on a subset of participants from the UK Biobank between 2006 and 2010. The land use regression (LUR) model was constructed to calculate the concentrations of particulate matter (PM2.5, PM2.5-10 and PM10), nitrogen oxides (NOx), and nitrogen dioxide (NO2). Binary logistic/multivariate linear regression models were applied to explore the potential linear relationships between air pollution exposure and newborn low birth weight (LBW) or BW. The Cochran-Armitage trend test was used to explore the possible association between the air pollution level and LBW. A restricted cubic spline (RCS) transformation of exposure variables was applied to visualize the relation of air pollutants to BW. Exposure to air pollutants, especially PM2.5 and PM10, was positively associated with LBW, and the odds ratios (ORs) and 95% confidence intervals (CIs) for each 10-μg/m3 increase in PM2.5 and PM10 were 1.25 ([1.03, 1.51], P = 0.025) and 1.12 ([1.02, 1.24], P = 0.021), respectively. A negative correlation was observed between the BW and PM2.5 (−0.05 [-0.08, −0.02], P = 0.001), PM10 (−0.03 [-0.05, −0.02], P < 0.001), PM2.5-10 (−0.04 [-0.07, −0.01], P < 0.001) and NOx (0.00 [0.00, 0.00], P = 0.021). Additionally, the BW changed dramatically up to a specific point (PM2.5 for 10.74 μg/m3, Pnonlinearity = 0.004; PM10 for 16.06 μg/m3, Pnonlinearity = 0.004; NO2 for 25.58 μg/m3, Pnonlinearity <0.001; and NOx for 39.88 μg/m3, Pnonlinearity <0.001), subsequently becoming relatively stable. PM2.5 and PM10 exposure were positively associated with LBW, and a negative correlation was observed between PM2.5, PM2.5-10, PM10 and NOx and BW.