The impact of patent traps and following litigations on technology innovation has become increasingly prominent in recent years. Patent early-warning is an important method to coping with the situation. As the first published patent documentation, patent applications can be used as a breakthrough. By analyzing feature fields in patent applications, a regression model of application features and patent litigations is established. The results show that the probability of patent litigations has significant positive correlation with the number of claims, the number of citing patents and the width of technology, and they are key features closely related to patent litigations. Thus the regression model can be used as a quantitative analysis tool in targeted technology fields to screen out high-risk patents easily leading to litigations, and conduct patent early-warning by using the first published application information.