Abstract: Extensive pesticide application can lead to the residues of harmful chemicals, such as highly toxic pesticides and heavy metals in agricultural products. The traditional air-assisted spraying has posed a great risk to human health in environmental ecosystems, particularly for the pesticide deposition and drift of off-target areas. Orchard target-oriented spraying technologies have been used to reduce the deposition and drift caused by off-target spraying, in order to control the environmental pollution within an acceptable range. In this study, a flow Pulse Width Modulation (PWM) model in a single nozzle was established for an orchard tower air-assisted sprayer using the grid volume calculation of the fruit tree canopy. The sprayer was equipped with 10 nozzles on each side. A Light Detection and Ranging (LiDAR) was selected to divide the fruit tree canopy into 10 grids in the vertical direction, where the nozzles had corresponded to the grids one by one. The spray volume was adjusted to control the flow of a single nozzle in different areas of the fruit tree canopy. The adjustment range of the optimal PWM duty cycle was 0-60%. The stability time of pressure PID control in the spray system was less than 3 s, and the pressure control deviation was less than 0.15 MPa. The opening and closing time of the nozzles were obtained through high-speed photography, and the delay compensation distance was determined to be 96 mm at a speed of 1 m/s. A Controller Area Network (CAN) bus communication protocol was then established, according to the spraying requirements in an orchard. A target-oriented variable-rate spraying control system was developed to integrate the target-oriented spraying control with the orchard sprayer. A physical prototype was also prepared in the target-oriented variable-rate sprayer. The laboratory and orchard tests were designed to evaluate the performance of the target-oriented variable-rate spraying control system. A stepped calibration plate was selected to simulate the target in the lab test. The width of each step of the calibration plate was 70 mm. The horizontal distance errors for the ahead or lag of the target-oriented spraying were obtained to compare the starting position of the spraying and the corresponding step position of the calibration plate. The test results showed that: 1) The optimal grid width was 210 mm at a speed of 1 m/s. The opening and closing lag distances of the nozzles were 19, and 41 mm, respectively. An evaluation was proposed to verify the target-oriented control accuracy in the orchard test, where a white cloth was used to record the spray range. The horizontal distance errors of the ahead or lag of the target-oriented spraying were obtained to compare the spray width on the white cloth and fruit tree canopy. It was found that the opening and closing lag distances of the nozzles were 122, and 185 mm, respectively. 2) A target-oriented variable-rate spraying test was carried out in a 5-year-old peach orchard. The continuous and variable-rate spraying was evaluated to arrange the water-sensitive measurement at the up, middle, down, left, and right positions on the orchard canopy. The spraying volume was recorded in real time. The spray application rate was evaluated to compare the continuous spraying, according to the spraying volume. It was found that the spray coverage of the target-oriented variable-rate spraying was less than 30% under the spray deposit densities greater than 20 deposits/cm2, which was much lower than the defined threshold for the overspray. The spray volumes were 4.53 and 1.71 L, respectively, for the constant-rate and target-oriented variable-rate spraying in the test area, indicating 62.25% less spray volume. As such, the target-oriented variable-rate spraying can be widely expected to fully realize the requirements in an orchard, in terms of the canopy position and volume changes of the fruit tree. This finding can also be used to rapidly promote the application of precision spraying technology in orchards.