Text steganography has become a hotspot in the field of information security in recent years. As an important steganographic carrier, text does not have the advantage that other multimedia could easily use more redundancy to hide information. Traditional methods were realized by changing the appearance attributes of text or by equivalent replacement based on grammar and semantics, but there were some shortcomings more or less. Based on Markov chain model, this paper proposed a coverless text steganography method. In order to enhance the effect of steganography, the method focused on the most important concepts in Markov model, the transition probability, and guided the generation of steganographic text based on it. The relevant experiments were designed to evaluate the model. The experimental results showed that the model had higher concealment and hiding capacity.