Abstract: In order to effectively measure the similarity between different probabilistic linguistic term elements，a similarity of probabilistic linguistic term elements considering the shape and position of data is proposed. Firstly，by comparing the shapes of the scatter plots of different probabilistic linguistic term elements，the probabilistic linguistic term shape similarity is proposed. Secondly，based on the earth mover's distance(EMD) and the completeness of the information，the distance similarity of the probabilistic linguistic term element is defined. Then，in order to measure the overall similarity of different probabilistic linguistic terms，a comprehensive similarity measure of probabilistic linguistic term element considering shape-position is proposed. Finally，a multi-attribute decision-making model is established based on improved radar graph，comprehensive similarity measure of probabilistic linguistic term element and max-min operator，and the rationality and effectiveness of the proposed method are illustrated by examples.

Key words: probabilistic linguistic term element, similarity measure, earth mover’s distance, max-min operator, improved radar graph, multi-attribute decision making