Compared with the traditional load forecasting,the spatial load forecasting pays more attention to the load distribution in a certain space, so it can better determine the selection and spatial layout of the electrical equipment. The rapid development of distributed energy and electric vehicles makes the urban spatial load distribution more complex. The original load forecasting method based on time series may bring large error, which is not conducive to the economy and reliability of urban power grid planning.Due to the nonlinear mapping ability of least squares support vector machine, a spatial load forecasting model for distributed and electric vehicle charging load is established. Finally, a practical example in a certain area of central China shows the effectiveness of the proposed method.