Motor imagery (MI), i. e., the mental simulation of an action without its actual execution, is a promising technique to boost motor learning via physical practice in rehabilitation, sport, and educational fields. The purpose of the present pilot study was to test the feasibility and the effectiveness of the application of MI as learning methodology place alongside conventional teaching technique as employed for physical education lessons. Thirty-three high school students from two classes were enrolled for instruction in the underhand serve in volleyball. One group, the motor imagery group (MIG) carried out the physical exercise along with the kinesthetic MI of the action, while the other group (the control group) was limited to the merely physical exercise. The training period lasted 8 weeks. MI duration and the duration of real movement (ME), the isochrony index (differences between real and imagined movements duration), and the number of balls which passed over the net (NBN) were evaluated before and after training. Results showed a significant improvement in the isochrony index for the MIG group exclusively; namely, MI duration became more similar to ME duration. Moreover, in MIG a significantly negative relationship appeared between the percentage change in the isochrony index and the difference between NBN before and after training. These findings suggest improvement in sensorimotor representation of the action, which lies at the basis of enhanced motor performance. The present study constitutes initial proof of concept on the application of MI as learning technique applicable to physical education lesson at high school.