The research study presents a real time method based on some video and image processing algorithms for eye blink detection to voice conversion. The motivation of this study is to help the disabled people who cannot communicate with human. The Haar Cascade Classifier is used for face and eye recognition to gain details about the eyes and facial axes. In comparison, the same classifier is used to assess the relation among the senses like eye and axis of the face to position the eye based on haar like features. This proposes an effective eye detection system which uses the sensed face location. An efficient eye detection method is proposed which uses the position of detected face. Finally, an eye blink detection method is done based on eyelids movement whether it is open or closed and is used for controlling mobile phones. The study have designed a very low-priced device that coverts the eye blinks to voice message with more accuracy compared to existing system. The eye blinks that are detected can be helpful in applications such as health assistance, S.O.S, basic utility. Test results show that our proposed system for a distance of 35cm delivers a overall accuracy of 98% and a detection accuracy of 98%