Instructional media especially audio visual media is very lacking in science learning causing students to get low learning outcomes. This study aims to determine the effect of the use of audio visual media on science learning outcomes of fourth grade students of SDN Duri Kepa 05. This study used a pre-experimental design type one group pretest-posttest design. The sampling technique that will be used by researchers is the nonprobability sampling Purposive sampling model. The number of samples used were 28 students from 31 students of class IV C. Instrument research was a multiple choice test of 30 questions and questionnaires as many as 20 statements. Before the instrument was used first it was tested on 28 students of class VB in SDN Duri Kepa 05. The results showed that the average pretest value was 59.29 and the posttest value was 75.07, for the significance value at the pretest of 0.126, the significance for posttest is 0.082, and the significance of audio-visual media is 0.200, each significant value of more than 0.05. Hypothesis testing with t test obtained a significant value of 0,000 <0,05, which means that there are significant differences, while the value of t is equal to 12,515 which means that the pretest value is smaller than the posttest of 12,515 so it can be stated that audio visual media have positive significant effect on learning outcomes.