The purpose of this study was to investigate the effects of the Assurance Relevance Interest Assessment Satisfaction (ARIAS) strategies set with Numbered Heads Together (NHT) on students' learning motivation. This quasi experimental study employed a pretest-post test nonequivalent control group design. The population of the research was all the eleventh graders who were enrolled in the science classes of public senior high schools (SMA) in Makassar, Indonesia. There were 21 schools in total. Random sampling technique was used to select 78 participants who were divided into two groups of treatment: 39 students for each treatment. Research instruments were developed based on the Attention, Relevance, Confident, Satisfaction (ARCS) model. Data on students' learning motivation was collected using a questionnaire that was distributed before and after learning. Research data was analyzed using a descriptive statistic and the analysis of covariate (ANCOVA) statistic. The results of the research suggest that learning strategies have an effect on students' motivation. Compared to ARIAS direct learning model, NHT-ARIAS is more potential to improve students' motivation.