Abstract: At present, stability of ozone forecasting products provided by numerical model and statistical model in Shaanxi Province was not well enough to meet the needs of refined environmental management. It is urgent to develop a practical and local ozone forecasting method in Shaanxi Province. In this study, the ozone prediction workflow is constructed and the relevant technical parameters and requirements are defined. Taking Shaanxi Province and Xi'an as examples, we analyzed the historical meteorological and ambient air quality data to obtain the ozone pollution pattern of Shaanxi Province, the distribution characteristic of ozone level under different meteorological conditions in Xi'an and the key points of typical case prediction. The relevant requirements for the construction, consultation and prediction review mechanism of regular case meetings are clarified, which provides a scientific and technical route for local ozone prediction model. In this study, ozone concentration has a significant negative correlation with relative humidity, air pressure and wind speed, and a significant positive correlation with temperature. The ozone prediction should focus on the ozone pollution level under the meteorological conditions of daily average temperature higher than 28℃, daily maximum temperature greater than 34℃, relative humidity less than 57%, air pressure less than 959 hPa and wind speed less than 2.3m/s. The accuracy of ozone prediction under this research method is significantly improved, which provides a technical reference for ozone prediction in all provinces of the country.