The outbreak of New Coronary Pneumonia COVID-19 and its spread throughout the China and many foreign countries have produced a huge social impact. Studying the dynamic characteristics of epidemic transmission will help us better control and prevent the epidemic. We have developed a discrete stochastic method to simulate the evolvement of the epidemic in Hubei Province of China. Firstly, the daily confirmed number of patients was processed according to the Erlang probability distribution of the queuing theory, and the daily number of patients onset and infected were obtained. The results are compared with references recently published by Chinese CDC, validating the scientific credibility of the method. Then, the effective reproduction rates at different stages of the epidemic are inverted to fit the number of daily onsets, and predict the future trend of the epidemic. It was found that the basic reproduction number R 0 decreased from 6.1 to 4.0 at the initial stage of the epidemic. After taking dramatic measures to close the Wuhan city, the effective R value decreased below 1, and gradually decreased to below 0.13. The peak of the onset of patients has already passed in early February. Although there small fluctuations in the epidemic are not ruled out, the overall trend will not change as long as strict quarantine measures are adhered to. The epidemic is expected to end around the end of March and the cumulative number of patients will reach around 71,000. Migrant workers and students returning after the Spring Festival are unlikely to induce a large epidemic rebound. However, some countries in the world are at the stage of possible outbreaks, and China should pay attention to inspecting and quarantine international travelers.