Billiards is a sports of a high entry barrier. According to field observations, it is difficult for novice players to learn and practice billiards techniques independently, for they struggle to imagine ball collision states and consequently find it challenging to perceive aiming directions. For this, we propose a holographic billiards learning assistance system based on Hololens 2. The system calculates shot selection and cue ball control strategies, displays the aiming position as a holographic ball in the user’s field of view, and presents cue ball control strategies. The implementation process consists of three steps: using a YOLO1 model to detect and locate billiard balls in images captured by loacatable camera, reconstructing the distribution of all balls on the table based on camera positioning data; calculating the optimal shot selection and aiming position, and inputting the table distribution into a pre-trained FCNN, which computes the optimal cue ball control strategy for the current moment; locating holographic balls at calculated aiming points to overlay with reality, and displaying cue ball control strategies. We conducted experiments with seven participants on a real billiard table and statistically validated that the system effectively guides players in aiming and executing proper cue ball control operations.