The changes in power load patterns and distribution locations pose significant challenges to the stability and long-term development of the smart grid system. Therefore, it is of great significance for power grid planning. The rationality of grid planning directly affects the stability and reliability of the smart grid system. This paper proposes a smart grid planning data decision scoring model based on random forest (RF-ETOPSIS). Firstly, a scoring mathematical model based on controller sorting was designed to score the grid planning data. Afterwards, a data decision-making method for grid planning based on adaptive random forest was proposed, which improves the accuracy of decision-making by constructing multiple decision trees. In the experimental phase, the proposed model shows 24.7% higher average performance than related models.