Abstract: Based on the CMA-MESO (China Meteorological Administration MESOscale weather forecast system) horizontal 3 km resolution 3 h cycle rapid update assimilation and forecast system, an hourly cycle analysis and forecast system was established. In this system, the background error correlation structure is improved by adopting a Gaussian correlation model with five-scale superimposition and introducing an anisotropic correlation scale scheme, and the impact of introducing a global large-scale information scheme on the analysis and forecast of the hourly cycle is examined. Numerical simulations of the strong convective case in eastern China on July 19, 2020, show that: (1) The hourly cycle absorbs more high-frequency observations and uses the more proximate 1 h forecast field as the background field in the cycle, which improves the quality of analysis and short-range forecasts compared to the 3 h cycle. (2) The introduction of large-scale information from the global forecast field to the hourly cycle regional analysis can weaken the influence of regional observations, which can negatively affect forecasting. (3) The improved five-scale superimposed Gaussian correlation model and the anisotropic horizontal correlation scale make the representation of the background error horizontal correlation coefficients of wind variables closer to the statistical results of the samples. Thus, the analyzed wind fields are closer to the observations in the hourly cycle, and the composite reflectivity and short-term precipitation forecasting of the strong convective process in Eastern China are closer to the real situation.