The measurement of textual complexity is a fundamental issue on e-learning, online question-answering, web search and browsing, etc. But it is still a challenge because the measuring process includes a series of cognitive mechanisms. According to the properties of cognitive principles in human text reading process, the different derivatives of textual keywords are given. Based on the categorical invariance in human conception learning, textual invariance is defined, and then the measuring method of textual complexity is developed which is consistent with the view of categorical invariance in human concept learning. Experiments prove that there is a strong correlation between the text understanding difficulty given by human and the textual complexity measured by our method. That is to say, the proposed method can reflect human's reading experience to some extent.