本論文主要目的是讓撞球機器人能夠具備接近人類撞球比賽的思考決策，使得撞球機器人能夠判斷局勢來決定攻擊或防守，並針對防守決策讓母球將目標球撞至對手不易進球之相對位置，找出母球各種移動路徑以及停留位置，決定最適合打擊方式(桿法)及力道指標，讓撞球機器人能有「守」有「攻」，因而更接近撞球選手比賽時的思考，達到贏得比賽之目的。 由攝影機擷取影像，找出各球之顏色及球心，再以所撰寫之VB程式判斷有無障礙球，撞球機器人可判斷母球、目標球及球袋之距離與角度關係，利用過去設計之灰色攻擊決策，決定攻擊桿法，以利於攻擊目標球，提升進球之成功率，但要讓系統貼近人類之思考方式及更具有智慧型判斷，於是加入灰色防守決策，利用碰撞理論，找出母球在各種力道撞擊後所停留的位置，透過灰決策，選擇最適合防守擊球的力道、角度及擊球位置，最後以灰決策整合撞球機器人之防守與攻擊，建立完整的擊球決策系統。 The objective of this thesis is to design a billiard robot possess with the ability of strategy thinking. It is just like human beings to play billiard games. The billiard robot is able to judge the situations for making offensive or defensive decision intelligently. The purpose of defensive strategy is to reduce the opponent’s chance to score and to enhance opponent’s mistakes. The various shooting strength and stroke strategy are decided by the predicted moving routes and the final position of the cue ball. The final goal of this robot is to win the billiard game like human beings does. In the research, a CCD camera is applied to capture the image of all the balls by the designed VB program in order to find out whether there is a block ball on the straight line between the cue ball and corresponding pocket. Then the offensive strategy is decided by three parameters in the Grey theory. Those parameters are distance between the cue ball and the objective ball, the distance between the objective ball and the corresponding pocket, and the corresponding shooting angle. By the way, the defensive shot is decided by the collision theory and the Grey theory from the predicted parameters which are the moving routes and the final position of the cue ball. Finally, the defensive and offensive decision-making subsystems are integrated into the billiard robot. The calculated results are shown in the VB interface to implement the shooting strength and stroke strategy. Experimental results show that this billiard robot are able to make a defensive or an offensive shot intelligently and successfully.