China's power industry is promoting reforms in energy types and data processing. The construction of power transmission and transformation projects, which occupy an important position in the power construction industry, has a high degree of cost risk due to its wide scope and long duration, which can easily have adverse effects on the construction of the power industry. Therefore, it is necessary to evaluate the cost risk of power transmission and transformation projects. On the basis of reading relevant literature, establish a cost risk set for power transmission and transformation engineering, analyze and screen out 15 influencing factors, and use the Extreme Learning Machine (ELM) algorithm of Genetic Algorithm (GA) to train the prediction model. After verification, the model accuracy reached 0.99874, indicating that the model established in the study meets the prediction needs. Finally, using this model to analyze risk assessment examples, Monte Carlo simulation (MCS) was introduced, and it was found that the excess risk of the project cost reached 59.94%, indicating a high risk. Corresponding measures should be taken to control it. Research has shown that the establishment of the MCS-GA-ELM model can play an auxiliary role in predicting and controlling the cost risk assessment of power transmission and transformation projects, providing a new approach for cost risk assessment of power transmission and transformation projects.