新藥的開發平均約需12-15年，其所耗費的金額相當可觀，隨著科技的進步，電腦模擬技術與藥物化學學科的結合，產生了電腦輔助藥物設計的新學門，其主要目的在於縮短藥物研發時間，而最終目標則是解決人類的疾病問題。 本研究由藥物設計的策略與方向進行說明，並將電腦輔助藥物設計依據其不同的理論基礎分為兩種，一為間接藥物設計方法，另一為直接藥物設計方法；文中內容將不同的軟體所提供的訊息與結果作一說明，透過藥效基團模擬及分子對接模擬的方式，來說明藥物分子和藥物作用標靶的結合情況。 論文的研究中透過間接藥物設計方法－藥效基團模擬，針對酪胺酸激酶抑制劑來進行分析與探討，得到一系列分子中重要的化學結構特徵，再透過測驗組的活性測試對於藥效基團模型的Hypo1來進行檢驗，以降低計算上的盲點。然後藉由直接藥物設計－分子對接模擬，我們可以實際探討受質和分子之間的結合情形，並且以小分子經由藥效基團模擬所推測的環境和大分子的模擬計算來做互補之討論，並於最後提出活性更佳的潛力藥物分子。 Quantitative structure–activity relationship (QSAR) methods have been demonstrated as an effective tool in discovering novel lead compounds. And pharmacophore modeling is one of the 3D-QSAR methods which can used to approach for generate chemical features. This method is based on the 3D structure information of molecules, and has been successfully applied to the drug discovery. In an effort to establish a pharmacophore model for EGFR and HER2 inhibitor that could serve as a guide for the rational design of further potent and selective inhibitors. We have developed a quantitatively predictive chemical function-based pharmacophore model by using the HypoGen algorithm implemented in the Discovery Studio 2.1 software. The most optimal hypothesis for EGFR model consists of three features: one hydrophobic (HYD), one hydrogen bond acceptor (HBA), and two ring aromatic (RA) function. And for HER2 model consists of three features: one aromatic hydrophobic (aroHYD), one ring aromatic (RA), and two hydrogen bond acceptor (HBA) function. To further validate our design rationale, protein-ligand docking software was used to elucidated the intra-molecular interaction. According to the docking result with FMM, O3P, and quinazoline type inhibitors, the ligand position relative to the adjacent residues was further constructed. Therefore, the built pharmacophore model could help us to better understand how the substituents influence the activity and afford important information for both indirect and direct drug design.