PurposeThe purpose of the study is to investigate the effectiveness of applying the Flipped Classroom Instructional Model to the rural online tutoring program for Upper Elementary Mathematics. Design/methodology/approachThe study took the unequal-group pretest-posttest quasi- experimental design to conduct the teaching experiment. The research setting was the Digital Partner Online Tutoring Program of National Dong-Hwa University, which was supported by the Ministry of Education of Taiwan government. The participants consist of 54 upper grade students from rural elementary schools, with 25 students in the experimental group and 29 students in the control group. The effectiveness of the experiment was detected by the factorial design of covariance with pretest as covariate and grade level as moderator. FindingsThe results of the study indicated that the experimental group’s overall learning performance on posttests was significantly better than the control group’s after the effect of pretests had been controlled. And the differences all achieved medium effect size. Among the four mathematical concepts, the differences on the two more advanced ones, “Common Factor” and “Common Multiple” even achieved large effect size. Originality/valueThe application of “Flipped classroom model” in various fields provides a prospective possibility for teaching innovation. Its critical factors, “using technology” and “participatory interaction”, exactly match the characteristics of the on-line tutoring program, including the on-line learning environment and the one-on-one interaction. The study provides empirical evidence for the effectiveness of applying the Flipped Classroom Model to the online tutoring program. Keywords: Upper Elementary Mathematics, Remedial Instruction, On-line Tutoring, Flipped Classroom Instructional Model