Promoter identification is an essential task in the research of transcription regulation, but computational prediction of promoters has been one of the most elusive problems despite considerable effort devoted to the study. A new prediction algorithm based on the combined features for E. coli σ⁷⁰promoters is proposed. According to their location, all promoters can be classified into two classes: promoters in non-coding regions and promoters in gene regions, and will be processed respectively. In each region, the features of primary sequence，including 1 content feature、5 signal features and 4 structure features, are combined and defined as a 10 dimensional vector, then the vector of combined features is further used by quadratic discriminant analysis to predict the potential promoter regions. The algorithm has been trained and tested on E.coli σ⁷⁰ promoter dataset by the jackknife method. The average prediction accuracies for “non-coding” promoters and “coding” promoters are 86.7% and 82.4%, respectively. The results indicate that our algorithm outperforms most of the existing approaches based on several performance measurements. Furthermore, algorithm framework is extendable and can accept more new features to improve the prediction results efficiently.