Abstract: Abstract: During the process of growing and selecting corps, it is significant to high-frequently gain of biological indexes of individual samples and their surrounding environmental parameters. Under normal circumstances, the number of samples is large, data acquisition cycle is long. To realize high-level automation of the process of growing and selecting crops, Data collection system (DCS) of corps was designed based on micro automated guided vehicle (AGV) in this paper. Techniques of advanced RISC machines (ARM), radio frequency identification (RFID), sensors, wireless communication, and modern control, etc were also use to the DCS. The DCS consisted of micro AGV, VDAS (vehicle data acquisition system), communication and control system. Specifically, the micro AGV, made up of control unit, action unit, guiding unit and orientation unit, was used to automatic navigate and pinpoint the location of samples. S3C6410 chip was use as the core processor of the control unit in micro AGV, S3C6410 is common RSIC processor developed by Samsung Company based on ARM1176JZF-S core and 16/32, which met the data processing requirements. ASLONG GA20Y180 micro direct current motor was used as the drive of the action unit, and achieved control of the motor L293D-based control module. Optical guided navigation was used to the guiding unit, which achieved reliable navigation through two micro AGV navigation modules. By RFID and optical recognition two kinds of ways, the orientation unit achieved targeting and accurate positioning of the Micro AGV during movement. The VDAS, made up of data acquisition units of image and environment as well as data processing unit, was used to collect data of samples' images, environmental humidity and temperature, carbon dioxide intensity, illumination intensity, and then to process and store the collected data. The communication and control system, made up of vehicle communication unit, and control software on remote control computer, was used to realize long distance transmission and control. When collecting the sample's data, the control software sent orders and the micro AGV carrying VDAS began to collect images and environmental parameters according to the planned routine. In order to validate the accuracy and stability of the DCS, taking soybean pot as sample in this paper, experiments on image and environmental data acquisition was done. It turned out that the images obtained from the DSC were evenly in good quality which met the requirements of image processing in the later period. Besides, the errors between the automatically collected environmental data and manual data were at around 2%, which met the precision standards of data acquisition. The DCS operated stably during the experiments and phenomenon of out of routine didn't occur. The error of orientation was fewer than 6 mm. It took the DSC 9 minutes to collect images of 160 samples, which demonstrated that the efficiency was improved greatly. This paper overcame the problem of data acquisition of individual samples when growing and selecting corps. It provides a good reference for the automatic acquisition of greenhouse corps.