The present study examined whether instructions emphasizing the relationship between science and mathematics would improve students' learning of science. Students in four classes of public junior high schools received two types of instructions concerning objects with changing velocity. Students under the experimental condition learned the relations between time and velocity (or distance) for objects moving in different conditions first, and then designed experiments for themselves to obtain a particular time when an accelerating object would reach a particular velocity. On the other hand, students under the control condition examined objects moving in different conditions in terms of time, velocity, and distance, and deduced the relations between these values. It was found that students under the experimental condition did better on tests about objects with changing velocity than students under the control condition, and reported that they were more interested and eager in the studying.