Genetic neural network model was used to analyze the intrinsic relationship between average ecological footprint and its influencing factors each year in Chengdu, and the developing trend of ecological footprint in the future was predicted. The average ecological footprint in Chengdu from 1985 to 2005 was increased slowly from 1.344hm2 to 1789hm2. Then 6 influence factors were selected to establish BP artificial neural network, which was made up of one input layer of 6 inputs, one output layer and one hidden layer. The global optimization performance of Genetic Algorithm was used to carry out the initial weights optimization of BP artificial neural network, so as to eliminate the flaws which can easily lead BP artificial neural network into local minima. The fitting accuracy of genetic neural network model to learning samples (average ecological footprint from 1985 to 2002) reached 99.70%; the simulation accuracy to testing samples (average ecological footprint from 2003 to 2005) reached 99.10%, which was higher than ordinary BP neural network (97.89%), and showed the high application value of the model.The predicted values of each influence factors from 1985 to 2005 were used as network input indexes, and the optimized network was used to predict the ecological footprint in the following years in Chengdu. The predicted values of average ecological footprint in Chengdu from 2008 to 2010 were 1.939 hm2, 1.990 hm2 and 2.049 hm2, and average ecological deficit were 1.629 hm2, 1.688 hm2 and 1.749 hm2. In order to alleviate environment pressure caused by consumption of urban development, excessive population growth in Chengdu should be controlled effectively, and the concept and structure of consumption should be renovated.