The touch input is generally used to operate smartphones. However, there are many occasions when an alternative to touch input is needed, such as when both hands are occupied or when fingers are out of reach due to one-handed operation. The eye gaze input is an alternative to touch input. The eye gaze input method generally uses the gazing point detection method, which detects where the user is looking at on the screen. However, gazing point detection requires an advanced camera, which is expensive to implement and difficult to determine the intent of input. To solve these problems, Eye Glance input interfaces in smartphones have been studied. This allows input by moving the user’s eyes back and forth from the center of the screen to one of the four corners of the screen. This eye movement has clear input intent and allows for fast input. The discrimination of eye movement is performed using time series data of eye feature point coordinates using an image processing library. However, there are issues with erroneous discrimination of eye movements caused by camera shake and facial movements. In addition, the system has only four choices, which is insufficient for complex input such as character input. In this study, we will improve the accuracy of Eye Glance input and make it multi-selective. By using facial feature points other than eye feature point coordinates, we will create an algorithm to suppress false input due to camera shake and facial movements. In addition, by using facial movements associated with eye movement as input, the number of input operations can be increased, and multiple selection can be achieved. We will evaluate the usefulness of the proposed method by conducting input experiments.