Aiming at the difficulty of the intelligent forecasting and early warning caused by the startup of online monitoring of water and soil loss in the field of power transmission and transformation projects, a new intelligent forecasting method of water and soil loss in power transmission and transformation projects based on SSA-KELM was proposed. The sparrow search algorithm (SSA) was used to optimize the regularization coefficient and kernel function parameters of kernel-based extreme learning machine (KELM), and the rainfall environmental factor was used as the input of the sample to construct the SSA-KELM forecasting model of water and soil loss. The forecasting model was used to forecast the water and soil loss of a substation, and was compared with KELM forecasting model and support vector machine forecasting model. The proposed algorithm was tested for a long time with the water and soil conservation monitoring data from the self-developed field monitoring system. The forecasting results show that the forecasting model of water and soil loss based on SSA-KELM is effective and its forecasting accuracy is higher than other current forecasting methods.