The invention discloses an SVR-based method for estimating an effective wind speed of a low-wind-speed section of a wind power generating unit. The method comprises the two steps of SVR model training and model online use. In the SVR model training process, a sensor is used to obtain a training characteristic set and a target set; the characteristic set is normalized to obtain an SVR training set; a PSO algorithm is used to select a penalty parameter and a kernel function parameter; and a trained effective wind speed estimation model is obtained. In the model online use process, output data of the unit is obtained in real time and is normalized and input to a trained SVR model, and after the data passes through a low-pass filter, a final effective wind speed estimation value is obtained. According to the method, the output data of the unit is fully utilized; effective wind speed estimation can be performed for the wind power unit of the low-wind-speed section; the design process is simple and easy to implement; and the obtained effective wind speed estimation value can be used for increasing wind energy utilization rate of the unit and reducing mechanical loads of the unit and wind resource assessment of a wind power plant, so that the economic benefits of the wind power plant are improved.