Abstract: In order to seek for the green tide disaster's forecasting method that based on the dominant factor, on the basis of the development stage that relied on the average growth rate of Ulva prolifera green tide remote sensing coverage area data in 2014, analysed weight change of climate factors affecting the expansion of green tide through a support vector machine(SVM) weight calculation model, obtained the dominant factors' weight intersections and summarized their change regulation in different stages, used the monitoring data in 2015 to verify the result of this method, and the error is 1.8%～7%. The results showed that mean ratio of temperature and wave height in the early-development phase was below 7, in the rapid growth phase was of 4 to 6, in the stationary phase was of 5 to 7, and in the decline phase was more than 7. The difference of haze's average growth rate in the early-development phase was more than 2 or less than -2, in the rapid growth phase was of -3 to 2, in the stationary phase was of -1 to 1, and in the decline phase was close to 0.