Abstract: Information hiding technique is of great importance for network information security and how to search and detect secret messages transferred in network is crucial and practical to safeguard the national security. On Internet, a large number of steganographic programs use the least significant bit (LSB) embedding to hide message in digital images. Based on contaminated data analysis, a new steganalytic technique capable of a reliable detection of spatial LSB steganography is proposed and its mathematical model is built . This method can not only detect the existence of hidden messages embedded in images fast and reliably, but also estimate accurately the amount of hidden messages embedded by using a sequentially or randomly scattered algorithm.