The aims are to enhance the continuity and accuracy of surface ultraviolet radiation (UVR) data in smaller regions and mountainous areas, explore the spatiotemporal distribution patterns and variation characteristics of UVR under complex terrain, and establish a high-precision UVR simulation model, so as to provide a scientific basis for optimizing planting zoning, precise production management, and sustainable development of specialty agriculture for flue-cured tobacco. In this study, the temporal and spatial variation characteristics of ultraviolet radiation (UVR) were analyzed by the methods of climatic propensity rate and spatial statistical analysis with using daily observation data from eight tobacco ecological meteorological stations in Shiyan area from 2012 to 2023, and the prediction equations of UVR exposure were developed by ridge regression analysis. The results showed that the annual UVR exposure in Shiyan area generally distributed in accordance with the characteristic of three mountains surrounding two rives (Duhe River and Hanjiang River), showing the spatial distribution characterized by "more in the north while less in the south and more in the east while less in the west". The daily average UVR exposure was 0.54MJ/m², and the diurnal variation roughly showed a wave-like characteristic; the monthly variation of UVR exposure with the seasonal variation followed a distinct singel-peaked variation law, and the monthly average value ranged from 9.33 to 25.30 MJ/m²; the annual UVR exposure showed a clear upward trend, and the annual average value was 201.10 MJ/m². UVR exposure was positively correlated with daily mean temperature, daily maximum temperature, daily minimum temperature, daily difference in temperature and sunshine hours, while it was negatively correlated with relative humidity and precipitation. The annual daily scale and seasonal daily scale regression models of UVR were established by ridge regression method, and all models showed good fitting performance (RMSE < 0.2 MJ/m²). The model fitting results show that the autumn-winter season model (RMSE = 0.11 ~0.14 MJ/m²) is more advantageous than the year-round day model (RMSE = 0.17 MJ/m²), but the fitting effect of the two models in spring and summer (RMSE = 0.19 MJ/m²) is not as good as that of the year-round day model. Based on the dual effects of ultraviolet radiation (UVR) on flue-cured tobacco growth, the model-fitted UVR data can be utilized for targeted selection of suitable planting areas during the site selection phase. During the field growth stage, for regions with higher UVR levels or during periods of intense UVR (July-August), a comprehensive analysis of its impacts should be conducted. This analysis provides the scientific basis for implementing appropriate protective measures through either ecological adaptation strategies or production interventions.