This paper designs an automatic full-well test system based on PTC technique. This system is applied to measure the full-well and readout noise of CCD cameras to improve the development conditions for the imaging circuits design. Scientific CCDs usually have deeper full-well capacity and lower readout noise, making it suitable for measurement instruments with large dynamic range and high accuracy. The project mentioned in this paper is a satellite-born instrument measuring the components of atmosphere, which employs a spectrometer and a camera to capture the spectral image data of the atmosphere. Since the brightness of the target varies in a wide range, the measurement requires a large dynamic range of the camera. In addition, to obtain an adequate image spatial resolution when the camera is scanning along the trail of the satellite, the readout rate of the CCD is determined to be 2.5 MHz. E2V's scientific product CCD275 is selected to make the new camera. According to the datasheet of CCD275, it has a full-well capacity of over 700 ke^- and a maximum readout rate of 5 MHz. However, the maximum readout speed specified in the datasheet is not the speed, at which the CCD full-well performance can be guaranteed. The CCD factories usually measure the full-well capacity at a low readout speed, and obtain the maximum readout speed under the condition that imaging function is achieved. The full-well of a CCD is determined by the storage and transferring capability of the CCD. The charge transferring capability is affected by offset voltage, driving current, and clock phases, which are the design parameters of the imaging circuit in the camera. Theoretical calculations and design references can only provide rough ranges of these circuit parameters. In a low readout speed, these parameters do not significantly affect the full-well transferring capability. However, when the readout speed is increased close to a certain speed, the effect becomes significant and these parameters need to be finely adjusted to keep the full-well performance. Over the certain speed, however the parameters are adjusted, the CCD can not output full-well. The required speed of 2.5 MHz is the speed, at which the design parameters are needed to be finely adjusted for the best performant. When a parameter is changed, the full-well measurement on the camera is required to provide feedback. Due to the numbers of parameters and their complex relationship, during the development of the new camera, such adjustments and measurements are required in a large number of times. The process of full-well test with PTC method is quite complicated and each test also takes a lot of time. To improve the development condition, this paper investigates PTC technique and designs an automatic CCD full-well test system. The system includes a test bench and a software. The test bench is comprised of an integrating sphere, a dark box, the camera to be tested, the stands supporting all the devices, and a computer. The software implements the data acquisition and data processing. By adjusting a remote controlled current source, which drives halogen lamps in an integrating sphere, the software sets up flat-field light at different illumination levels. In each illuminance, the software acquires an image data from the camera. Each image contributes a point to the PTC, the horizontal ordinate is the average value of the image and the vertical ordinate is the variance. Densely acquiring the points of PTC will slow down the test, as every time of adjusting the light source takes a time to stay stable. On the other side, less points of the PTC will reduce the precision of the measurement, as break point of the curve, which is the full-well point, will be difficult to be located. This paper creates a novel PTC plot method called“quick-PTC-plotting”, to shorten the acquisition time without any decline of the precision. It firstly adjusts the light source with a large step and acquires corresponding images, roughly plotting a PTC with few points. Then it determines a narrow range, where the break point of PTC most likely exists, obtains images with a small step in the narrow range, and re-plots the PTC. The software implements the“Quick-PTC-plotting”and reduces the times of light adjusting. The software also acquires dark images and background images, which are used for the calculations of the PTC. The automatic CCD full-well test system has been applied in the real engineering. The methods and implementations have been verified. Statistic result shows the random error of the system is ±0.6%. Four space-grade CCDs, of which the full-well capacities have been tested by the factory with lower speed, are also tested in the system. The deviation of the tests in factory and the automatic system is less than 1%. The accuracy of the test system is adequate for the application of camera development. The automatic system shortens the full-well test time to 20 min, while manual operation takes 2 h. It also brings a benefit of unattended operation.