The invention discloses a height prediction method and system; the method comprises the following steps: A, obtaining juvenile basic data in an assigned age range and current height data, wherein thebasic data comprises yet not limited by the following items: sex, age, regions, body weight, diet habits, sleep habits, exercise habits and parent heights; B, selecting features of the obtained basicdata so as to extract key feature data remarkably affecting height; C, clustering and grouping the extracted key feature data and corresponding height data, and building a juvenile growth curve modelin different age ranges of different sexes in each group; D, selecting the growth curve model of the corresponding sex and age range in the corresponding group according to the feature data of a to-be-predicted height juvenile, and inputting the feature data into the selected model so as to obtain the juvenile height predicted value. The method and system can evaluate and parse the juvenile growthconditions in real time, can research and analyze the juvenile nutrition health status, and can make customized recommends for children diet life habits.