Geometric morphometrics is an important research method for the analysis of the cephalopod beak morphology, geometric morphometrics contains the landmark and outline method. Although the type of landmark is identified, the study of beaks based on the method of landmarks does not have a criterion for the selection of landmarks, so it is necessary to explore how many landmarks and samples are necessary to indicate the best variation of the beak pattern. Based on samples of 300 pairs of Uroteuthis edulis beak caught in 2018﹣2019, traditional morphological measurements of beak are used to explain basic morphological differences. Using the R-language geomorph package “gpagen” and the LaMBDA package “LaSEC” (Landmark Sampling Evaluation Curve) functions, the centroid size of the landmark data of 10, 20, 28 landmarks and 50, 80, 100, 200, 300 pairs of sample combination schemes is calculated, the centroid size is calculated, the landmark sampling evaluation curve is obtained, and the landmarks and sample numbers are explored to best demonstrate the degree of variation of the beak pattern. Finally, the morphological variation point of the whole beak sample is analyzed quantitatively, the point of large variation of the beak pattern is preserved, and the specific punctuation scheme is determined according to the best landmark number scheme. The standard deviation of the 12 morphological parameters of the beak between different samples was found to be large (Table 3), and the size of the beak increased significantly with the increase of the landmark (Fig. 3). Under different schemes, the slope and dispersion degree of the landmark sampling evaluation curve are significantly different, the three landmark schemes reach the optimal fitted value of the number of landmarks are different, 28 landmark scheme sampling curve with the increase of landmark collection subsampling curve slope gradually decreased, the variance gradually decreased (Fig. 4, 5). In order to best show the degree of morphological variation of the beak (F = 0.99), 26 landmarks were used in the upper and lower beak when the beak sample was about 50 pairs; The sample was 50﹣300 using 27 landmarks (Fig. 6). In addition, quantitative analysis of the morphological variation of the overall beak study sample (Fig. 7) found that, under the premise of ensuring the integrity of the beak form, 1﹣2 points were deleted on the basis of 28 landmark schemes to arrive at the best landmark scheme, that is, 26 landmarks: the upper beak deleted landmarks 5, 7, the beak deleted landmarks 5, 25; 27 Landmarks: upper and lower beak remove landmark 5. Using LaSEC method to quantitatively analyze the number of landmarks and samples, it is possible to propose a quantitative scheme to maximize the representation of morphological variation and centroid size. In addition, after analyzing the optimal number of landmarks, it is still necessary to combine the morphological variation map of the overall study sample, and according to the different types of landmarks definition, on the basis of ensuring the integrity of the pattern appropriate removal of small pattern variation points. Through a variety of considerations to clarify the beak landmark scheme, for the follow-up based on the landmark method of the head-foot beak morphology research to provide a scientific basis.