Ecological security is an important research topic, especially urban ecological security. As highly populated eco-systems, cities always have more fragile ecological environments. However, most of the research on urban ecological security in literature has focused on evaluating current or past status of the ecological environment. Very little literature has carried out simulation or prediction of future ecological security. In addition, there is even less literature exploring the urban ecological environment at a fine scale. To fill-in the literature gap, in this study we simulated and predicted urban ecological security at a fine scale (district level) using an improved Cellular Automata (CA) approach. First we used the pressure-state-response (PSR) method based on grid-scale data to evaluate urban ecological security. Then, based on the evaluation results, we imported the geographically weighted regression (GWR) concept into the CA model to simulate and predict urban ecological security. We applied the improved CA approach in a case study—simulating and predicting urban ecological security for the city of Wuhan in Central China. By comparing the simulated ecological security values from 2010 using the improved CA model to the actual ecological security values of 2010, we got a relatively high value of the kappa coefficient, which indicates that this CA model can simulate or predict well future development of ecological security in Wuhan. Based on the prediction results for 2020, we made some policy recommendations for each district in Wuhan.