目前利用網路進行測驗的情況雖然普遍，卻限於簡單的題型型態及較為靜態呈現方式，無法充分展現多媒體在網路上的應用能力。尤其是在九年一貫「自然與生活科技」的課程領域中，多媒體的測驗型式較能有效及精準的表達科學及科技測驗的意涵，並可避免因文字表達所造成的題意概念模糊，而使學生產生誤答。因此，本研究嘗試建置一個以題型模版為基礎的網路多媒體測驗編輯系統，內含多種測驗題型，並支援多種媒體格式，形成不同的試題編輯模版，以方便從事多媒體測驗的出題工作。本研究以系統發展生命週期進行系統開發，開發過程中藉由文獻探討進行系統需求分析、設計，並以問卷調查及準實驗研究等方式完成系統之內、外部評鑑。研究發現，專家對於本系統內部結構及行為大多持正面的看法，教師對系統的試用經驗調查亦屬正面的評價。而在學生使用本系統進行多媒體測驗方面，經由準實驗設計及共變數分析的結果顯示：藉由本系統進行多媒體測驗的實驗組學生，其測驗的平均成績高於一般傳統紙筆測驗的控制組學生，且達到統計上的顯著差異。這顯示運用本研究的測驗系統來實施科學及科技測驗時，學生會因多媒體元素的真實與完全表達題意，使得測驗成績較能反映學生本身的實際認知程度。

Common educational web-based assessment systems are facing two challenges in development: insufficient variety of question types and integration of media formats, as past literature stated. Especially in the field of science and living technology, multimedia-based assessment can more precisely and efficiently exhibit the core spirit of science and technology evaluation, and can avoid miscomprehension caused by simple text. Thus, this study tried to design and build up a web-based multimedia assessment authoring system featuring customized and easy to use templates – containing several question types and supporting a variety of media – to simplify and streamline the creation process. Therefore, instructors are able to create and manage a web-based assessment with pre-customized templates and media formats after lecturing; students are able to take the exam for self-evaluation and practice whenever learning activities are finished. This system was innovated and implemented through System Development Life Cycle (SDLC), and a Template-based Multimedia Assessment Authoring (TMA2) system was therefore formed, highlighting the multiple module function as well as the integration of various media types. Internal and external evaluations were conducted later on via experts’ comments. Survey results indicated that most of the experts unanimously held positive attitudes toward the system, and so were the instructors who actually implemented it in their lectures. With respect to students’ learning results, statistical data gained from quasi-experimental design and covariance analysis showed the experimental group with TMA2 support was more likely to achieve higher scores than the one without it; meanwhile, statistical significances were exerted, too. This result implies that with TMA2 support, learners will gain better comprehensive ability due to complete illustration of test through various media element, and will honestly evaluate learners’ cognitions. Finally, the author proposed further research implications and directions on this emerging topic.