The purpose of this study is to explore students' motivation towards science learning at different grade levels and to investigate whether inductive reasoning can contribute to an explanation of science motivation. The study conducted a cross-sectional assessment in six public schools in Vietnam with a total population of 813 students from the 5th, 7th, 9th, 10th and 11th grades. Students completed instruments in either paper-and-pencil or online administration modes. An adapted science motivation questionnaire comprised five subscales for self-efficacy, active learning strategies, science learning value, achievement goals and learning environment stimulation. An inductive reasoning test consisted of four subtests: figure series completion, figure analogies, number analogies and number series completion. The results of confirmatory factor analyses and Rasch model measurement showed that the instruments were adequate fit models, both the science motivation questionnaire (RMSEA=.054, CFI=.919, SRMR=.055) and inductive reasoning test (RMSEA=.038, CFI=.902, SRMR=.044). We found that students' scores gradually fell grade by grade in science motivation throughout the grade cohorts. Particularly, students' motivation dropped noticeably on the self-efficacy and active learning strategies subscales. No gender difference was found between males and females in science motivation. Although a positive correlation was observed between inductive reasoning and motivation across grade levels, multi-model Bayesian inference suggested that other factors, such as age, science performance and parental involvement, were better predictors of students' science motivation. Furthermore, a path analysis showed that inductive reasoning has an indirect effect on science motivation through a science performance variable. The implications for enhancing science motivation are also discussed.