The quality of education in Indonesia is still low based on data from international studies, namely the education development index shows that Indonesia has not moved from the medium or medium category, especially in the field of science. This fact assumes science learning as a problematic learning content. This study aims to develop a learning device that is able to overcome these problems on a local scale. This study uses the Design of Educational Research and its development model uses the Dick and Carey approach. The subjects used in the one-to-one test amounted to 3 people, the small group test amounted to 14 people, and the test of a large group of 60 class VII SMP students with a testing area in Banjarbaru. Learning outcomes show improvement in both the experiment and the control class with an increase category at the moderate level. Data shows normal distribution and both classes are considered homogeneous and the t-test shows that there is a significant influence between the use of guided inquiry models compared to conventional learning using developed tools.