現今大學林立，各種不同程度之學生均有機會攻讀學士學位，然而在教育過程中，對於影響學生學習成效的原因之注重程度又較以往式微的大環境下，如何能以學生目前學習方式、學習者性格、學習前測試結果、價値觀等因素去預測學生未來可能成績等第，進而對於教與學雙方都會產生若干預警的效能，如此對於提升教學成效也許會有所助益。本文針對技職學生，先以上述各因素進行網路問卷調查及基礎能力測試，再以混合系統對問卷及測試結果做推論。研究結果發現本學習成效預測系統對於所進行的部份專業科目成績預測效果佳，然而對於非專業科目之學習成效預測則稍有偏差。至於預測成績較佳或較差的同學均較爲正確，這對事前提出預警會有相當的助益。由於本研究所使用之各種不同程式語言所寫出之各部份的各項功能均爲自行研發設計者，所以對於未來研究之配合改進極爲方便。它的優點在於資訊的取得方便受測者，整個資訊處理過程全部電腦化，故不會有人爲因素之偏差，推論方式或統計計算均可以自行修正或選擇，不會有因須配合其他工業軟體之使用而造成困擾，對於結論的處理亦易於電腦化。另外，系統的個人化或個別科目局部化結果使得本系統發展具有相當的彈性。由於本次研究爲初創，對於還需加入預測系統之影響因素或其他推論方式將會於往後的研究中繼續努力發展。

Almost all students have opportunity to study in universities nowadays. However, there are less and less paying attentions on what factors affecting university students' learning achievement. It is beneficial for both teaching and learning if teachers can predict the students' possible future achievement by their learning manner, personality, value, and informal test results. In this article, we first use questionnaires to collect these factors through internet, and then we established a hybrid system to make an inference about students' future achievement by the factors. The research indicates that the predict results for curriculum achievement in professional subjects are much better than unprofessional ones. Since each specific function in the system is programmed with different computer language and designed by ourselves, it is flexible for future further researches. The advantages of the system are the convenience for collecting data, automatically information processes without any possible ignorance from people, statistic calculations and inference methods being chosen liberally, free from puzzles for using engineering software in some specific situations, and the output of the system being easy handled by computer. Besides, individualize of the system and localized of one specific subject make the system become rather flexible. There are more reasoning factors and other inference engines need to be added in this system for future study after this debut research.