Evaluation of facial movement, especially eyelid movement, has depended on the subjective judgment of trained clinicians. Recently, a few objective methods have been reported, but they required uncomfortable markers to be attached to the eyelids and a special-purpose, high-speed video camera. This study had two aims: one was to develop a new device for measuring eyelid motion dynamically and quantitatively, without eyelid markers or a high-speed camera; the other was to devise feasible parameters for eyelid motion. The system consisted of a personal computer with a general-purpose multimedia board and a software program that the authors named blepharokymography. A sequence of blinking eyes was recorded with a video camera. After the capturing process of the video, kymograms were produced from the movie file. Kymograms were converted to binary images by threshold filtering. The lower margin of the upper eyelid was traced, and displacement and velocity curves were obtained. Some parameters were devised and verified in preliminary clinical data. The analysis revealed that the displacement (8 mm in normal compared with 5.2 mm in paralysis), average closing velocity (74 mm s−1 in normal compared with 30.6 mm s−1 in paralysis) and peak closing velocity (154 mm s−1 in normal against 63.4 mm s−1 in paralysis) were useful parameters for differentiating the normal and facial-paralysis states.