To stimulate the accomplishment of provincial and municipal carbon emission targets, a Stochastic Impacts by Regression on Population, Affluence and Technology (STIRPAT) model is established based on the industrial carbon emission data of Linyi City from 2009 to 2019. The model quantitatively analyzes the relationship between industrial carbon emission and enterprises' fixed assets, per capita industrial production added value, energy intensity and energy structure in Linyi City. The collinearity between the variables is eliminated by ridge regression. For every 1% variation in enterprises' fixed assets, per capita industrial production added value, energy intensity, the proportion of raw coal in energy consumption and net purchased electricity will lead to 0.069 37%,0.016 30%,0.214 60%,0.550 00% and 0.214 60% fluctuation in the industrial carbon emission, respectively. The per capita industrial production added value is predicted by a Grey Model, GM(1,1). The industrial carbon emissions in Linyi City from 2020 to 2030 under six different prediction scenarios are analyzed. The comparison results conclude that maintaining a moderate growth of industrial product, increasingly optimizing energy structure and controlling energy intensity are the effective ways to reduce the industrial carbon emission of Linyi City.