Emerging trend detection is a new challenge and an attractive topic in text mining. Our research goal was to construct a model to detect emerging trends in a set of scientific articles; the resulting model is richer in topic representation and more appropriate for evaluating emerging trends than existing models. To achieve this end, we associated each topic with many features extracted from scientific articles and constructed two measures for ranking interest and utility. Based on the information commonly provided in scientific papers, our method can adapt to different kinds of scientific corpora and also can be efficiently modified to adapt it to user needs. We also built a prototype system to test the model and the evaluations show that our model promises to achieve significant results in emerging trend detection.