Concrete bridges, as critical components of highway construction, generate substantial carbon emissions during their construction phase, necessitating the development of a relatively precise carbon emission estimation model to promote low-carbon construction practices. The present study categorizes the sources of carbon emissions during the construction of highway concrete bridges into material production, transportation, off-site processing, and on-site construction. The carbon emission factor method is employed to calculate the carbon emissions during the construction period of 31 concrete bridges on a newly built expressway. An analysis of carbon emission characteristics and their correlations reveal that factors such as bridge length, total material weight, and machinery working hours significantly influence emissions during bridge construction. The Spearman correlation coefficients for these factors are 0.96, 0.88 and 0.82, respectively, with collinearity observed among them. Employing these variables, ridge regression, Lasso regression, and elastic net regression models were developed to mitigate collinearity. The Lasso regression model demonstrates the highest accuracy in estimating carbon emissions, with an R² of 0.976 2, thus making it the preferred model for estimating emissions during bridge construction. This model can calculate the carbon emissions for a variety of design and construction plans of concrete bridges based on bridge length and total material weight, serving as a methodological reference for the development of low-carbon designs and the optimization of carbon reduction strategies during the construction process.