The invention discloses a method and system for identifying key factor influencing power line waving, aiming at increasing precision of predicting power line icing and waving. The method includes: acquiring power line waving data and at least two corresponding potential key factors; establishing equations of linear regression between each potential key factor and power line waving, computing errors of every equation of linear regression, and determining an equation of linear regression that has the minimum error as a benchmark equation of regression; selecting a second potential key factor from the rest of potential key factors to adjust the benchmark equation of regression, and if the adjusted benchmark equation of regression has the error which is smaller than or equal to that of the benchmark equation of regression prior to adjustment, determining the adjusted equation of regression as the new benchmark equation of regression, and if the adjusted benchmark equation of regression has the error which is not smaller than or equal to that of the benchmark equation of regression prior to adjustment, providing no update; proceeding the rest in the same manner until all potential key factors are traversed, and the potential key factor contained in the finally obtained benchmark equation of regression being determined as the power line waving influence key factor.