Due to the complex composition of carbon emission driving factors of products/services, it is difficult to achieve carbon emission reduction effects. Therefore, this paper proposes a study on carbon emission reduction methods based on improved life cycle assessment algorithms and mixed integer linear programming. When using the mixed integer linear programming (MILP) model of linear approximation for specific emission reduction planning, define integer variables to represent whether a certain emission reduction technology is adopted, continuous variables to represent the resource usage and clean energy substitution at each stage, and the objective function of constructing the MILP model is to minimize the total carbon emissions throughout the life cycle. When the linear expression between the carbon emissions corresponding to the product/service and the nonlinear factors reaches convergence in the objective function, a carbon emission reduction plan is output. In this test result, the maximum reduction in carbon emissions during the life cycle reached 16.76%, and the overall carbon emissions dropped by 11.95%.