To assess whether Shaanxi Province’s transportation industry can achieve the carbon peak target by 2030 and to determine the total carbon emissions of the industry by that year. A model based on dual feature extraction (DFE) and an improved Pelican algorithm (IPOA) to optimize the kernel extreme learning machine (KELM) is proposed in this study and used to predict transportation industry carbon emissions from 2022 to 2040. First, the influencing factors of carbon emissions are extracted through Spearman and gradient boosting decision trees (GBDT), and the extracted factors are used as the input set of the carbon emission prediction model. Secondly, the IPOA is used to optimize the parameters of the KELM to overcome its shortcomings of easily falling into local optimal solutions. Finally, the combined model is used to predict the future carbon emissions of the transportation industry in Shaanxi Province. Comparing the prediction results and error indicators with other optimal benchmark models, the model improved by 19.98%, 30.72%, and 21.33% in the three indicators of MAPE, RMSE, and MAE, respectively. It is confirmed that the carbon emission prediction model proposed in this study is more effective and can more accurately reflect the future carbon emission trend of the transportation industry in Shaanxi Province, China.