This paper conducts a predictive study on the realization path of the “dual carbon” goals in the Inner Mongolia Autonomous Region. Based on the improved Logistic population prediction model (R2 ≈ 1), the GDP ridge regression model driven by macroeconomy and the second-order difference algorithm of carbon emission intensity, a coupled model was constructed. The dynamic coupling analysis of multiple driving factors was achieved for the first time. Research shows that: ① Carbon emissions in Inner Mongolia have significant phased characteristics. It declined during the 12th Five-Year Plan period, but increased by an average of 6.03% during the 13th Five-Year Plan period. ② Based on the historical data from 1999 to 2021, the model predicts that the carbon peak (peak of 104 million tons of CO₂ equivalent) will occur in 2031, which is one year later than the regional target. The target for achieving carbon neutrality is 2068, by which time the carbon intensity of GDP needs to be reduced to 50% of its peak. The research innovatively found that the contribution rate of energy efficiency improvement to emission reduction was 62.7%, which was much higher than that of population scale effect (9.3%) and economic scale effect (28.0%). The model demonstrates superior accuracy (SSE < 0.05, RMSE<0.12) compared to conventional intelligent algorithms. This achievement provides a quantitative basis for formulating differentiated policies in high-carbon emission regions. The proposed “population-GDP-energy” three-factor coupling model can provide a methodological reference for similar studies.