Abstract: Scenario generation forms a foundation for the application of sequential Monte Carlo method in the field of reliability assessment for power systems with renewable energy. Traditional time series generation methods cannot reflect both the seasonal and diurnal variations of wind-irradiation-load simultaneously, and few studies account for the temporal correlation of wind-irradiation-load at the same time. In this paper, we divide a year into twelve segments and extract the seasonal and daily characteristics of the historical data by month. These data are utilized to form a weather-load characteristic parameter library, and thus a time series generative models is established based on ITO random process. The aforementioned basis leads to the generation of a multi-source-load joint scenario. The simulation results demonstrate that the proposed time series generation method effectively captures the seasonal and diurnal characteristics of wind-irradiation-load. The generated multi-source-load combined scene has higher accuracy in reflecting the actual situation in a specific area.