The port scan is a well-known technique which malicious people often use before attacking a server. The attackers obtain the fingerprint of the target server by scanning ports and then make an attack scenario. Several approaches including the ‘port knocking’ and ‘Single Packet Authorization’ (SPA) have been developed to defense port scanning attack and allow only authenticated users to access ports. However, the approaches have a disadvantage that the attacker can obtain the information about the ports by applying inference techniques given observed patterns. If a router, connecting the server to the outside, is cracked by the attacker, he or she could infer particular ports which authenticated users consistently use to communicate with the server. In this paper, we propose a new defense method, Honeyport, which can prevent the attackers from obtaining the information about ports and make them demotivated by disguising the server as peripherals. Furthermore, by adopting packet encryption as in IPSec, the attacker cannot obtain the critical information via packet sniffing in our proposed model.