指出了當前反病毒軟件在檢測未知木馬方面的不足，介紹了人工免疫系統在反病毒軟件自適應性方面的優點，以及人工免疫機制在木馬檢測方面的可行性；通過對木馬新技術的分析，用一個木馬模型證明了現在計算機安全體系的不足，提出將木馬檢測從反病毒軟件中遷移到免疫型入侵檢測系統中作為子系統，利用其免疫機制來提高木馬檢測的自適應能力；并同時提出了依據進程的系統資源使用狀況來映射進程的系統調用的行為模式，以此建立了基于人工免疫機制的木馬檢測模型。

This paper points out the deficiency in detecting the unknown Trojan horse of the present anti-virus software at first, introduces the advantage of artificial immune system in self-adaptability aspect, and points out the feasibility of artificial immunity mechanism in Trojan horses detection; Then through an analysis about the new technology of Trojan horses, proves the deficiency of current computer security system with a Trojan horses model, presents the transfer of Trojan horses detection from the anti-virus software to the subsystem of immune IDS, improves the self-adaptive capacity of Trojan horses detection with its immune mechanism; Finally, a behavior mode is put forward, which is mapped from the using situation of process systematic resource to the process systematic call, and by this means, a Trojan horse detection model based on artificial immunity mechanism is set up.