The purpose of this study is to examine a method for quantitatively evaluating the lift-up effect of facial massage using beauty devices. It is said that facial massages have a lift-up effect that tightens the jawline and improves wrinkles and sagging. Currently, the lift-up effect is subjectively assessed through surveys, and there are few studies that quantitatively evaluate the changes in the skin caused by massage. In previous studies, aspects such as the jawline, angle of the chin, and face length and width have been quantitatively evaluated. However, there were issues such as the inability to evaluate multiple indicators simultaneously, and the time and cost involved. To address these issues, we focus on a machine learning model called FaceMesh, which can detect facial landmark coordinates in real time from images. By applying this model to evaluate the lift-up effect, it may be possible to perform objective analysis based on numerical data, without relying on traditional subjective evaluations. Moreover, since it can be evaluated in real-time using a webcam or smartphone camera, it can be easily used by anyone, allowing the effects of the massage to be visually confirmed in real-time. Therefore, in this study, the process of facial massage using a beauty device will be recorded with a webcam, and it will be examined whether the lift-up effect can be quantitatively evaluated from facial images before and after the massage using FaceMesh.