斜紋夜蛾 (Spodoptera litura (Fabricius))為食性廣且食量大之重要農業害蟲，過去常使用化學藥劑進行防治。但近年來，化學藥劑施用所帶來之環境污染及食用安全疑慮，漸漸受到民眾與政府重視，因而積極推廣非農藥防治。馬尼拉小繭蜂 (Snellenius manilae (Ashmead))為斜紋夜蛾之寄生性天敵，屬於單元性幼蟲內寄生蜂，寄主專一且分布廣泛，適合開發作為斜紋夜蛾生物防治資材。本研究即以建立馬尼拉小繭蜂大量飼養之系統為主，設計大量飼養之裝置及擬定適當之寄主-寄生蜂比例。實驗結果顯示，以幼蟲生長至一齡末時進行寄生蜂接種產繭量較高，但與二齡初寄主相比無顯著差異；寄生筒內所使用之初步飼材及收集蜂繭方式同時也會影響產量，以芋頭葉作為初步飼材顯著提升繭之產量 (與人工飼料飼養比較)，由於蜂繭停留在寄生筒內時間愈長，被其他個體破壞之機率愈高，因而測試了收繭方式對收繭量之影響，而結果顯示以連續分次收集之方式產量較高。寄主與寄生蜂密度關係實驗之結果顯示，以本實驗設計之寄生筒為前提，寄主-寄生蜂相對比例為100:3能得到60%收繭率，為本實驗中最高產量。此外，實驗結果也顯示斜紋夜蛾卵塊重與孵化蟲數成正相關，並求得迴歸線公式： 孵化蟲數= 22.39453×卵塊重-75.67616，R2=0.8513。未來可利用卵重及此線性公式推算接入之斜紋夜蛾數量，搭配適當數量寄生蜂進行接種，並於接種後12-15天收集蜂繭。期望此系統能便於未來相關實驗建立族群，以及提升馬尼拉小繭蜂應用於田間生物防治的可能性與潛力。 Spodoptera litura (Fabricius) is a serious pest for its wide host plant range and good appetite. Farmers used to control Sp. litura with chemical pesticide; however, the consequences of using pesticide, such as environmental pollution and food safety concern, had come to people and government’s attention. Therefore, government had been promoting non-pesticide control actively. Snellenius manilae (Ashmead) is a natural enemy of Sp. litura. It is an endo-parasitoid with strict host range, and widely distributed over Taiwan, hence to be considered as suitable biocontrol agent. The purpose of this study is to develop a mass-rearing system of Sn. manilae, including basic set-up of the parasitical column and suitable host-parasitoid ratio. The result shows that use last 1st instar as host produces more cocoons than early 2nd instar does, however there is no significant differences. Also, using taro leaves as initial feeding material and “continuous collection” produces more cocoons. In the experiment of host-parasitoid relative ratio, we determine that the best ratio to be used is 100:3 (host: parasitoid). Under the condition of parasitical column we designed, it produces 60% cocoons of host number in 12-15 days. Last, the relationship between Sp. litura egg mass weight (W) and hatched larvae number (N) was tested. Data showed that there is a positive correlation between egg mass weight and hatched larvae number, we also found equation of regression line (N= 22.39453×W-75.67616, R2=0.8513). We expect this research would bring contribution to correlative research, or increase the possibility of applying Sn. manilae into field as biocontrol agent.