本文讨论了一种新的基于特征状态的智能预估器。通过对典型化工大纯滞后对象的特性分析，提出了一种从对象控制作用和响应曲线中提取信息，构成特征状态，由此来驱动模糊规则的智能预估策略。它避开了传统的建模预估方式，更适于工程实际，其有效性通过实验得到了证实。 This article discusses a new intelligent predictor based on feature states. Based on the analysis of the characteristics of a typical large chemical pure lag object, an intelligent forecasting strategy is proposed, which extracts the information from the object control function and response curve to form the characteristic state and drives the fuzzy rule. It avoids the traditional mode of modeling and prediction, more suitable for engineering practice, its effectiveness has been confirmed by experiments.