Single micro-light emitting diodes(LEDs) with different sizes and array micro-LED are designed and prepared, where the sizes of the single micro-LEDs are in a range of 40−100 μm, their electrodes are all co-N electrodes, P electrode is drawn out alone; the number of array pixels is $ 8\times8 $ , which is a passively driving structure with a pixel size of 60 μm. In the process of device preparation, N electrode and P electrode are fabricated by the sputtering & stripping method. The electrode thickness is 2.4 μm. Thick photoresist 5120 is used as a mask, and N GaN is etched to the substrate by using the ICP dry etching to form an isolation trench. The PECVD technique is used to deposit an SiO ₂insulating layer with a thickness of 10000 Å. By optimizing the electrode structure and thickness, the reliability of the P electrode at the slope of the isolation trench is improved, and the SiO ₂insulating layer has good encapsulation; field programmable gate array (FPGA) is used to drive and display the micro-LED passive array. The single micro-LEDs of different sizes are tested and analyzed in the aspects of electrics, photics and thermotics and the results of which show that the current density corresponding to the peak radiation flux of 80 μm micro-LED is 1869.2 A/cm ², which is 57.1% higher than that of 100 μm micro-LED, indicating that the current density corresponding to the peak radiation flux of micro-LED increases as the size decreases; compared with the ordinary blue LED, the micro-LED has a large kfactor, and with the size decreases, the value of the kfactor increases, indicating that the micro-LED series resistance is larger, and the thermal stability is not so good as the traditional blue LED. Finally, the field programmable gate array (FPGA) can achieve a good drive for the micro-LED passive array. The driving principle is passive scanning driving, which is carried out in a row-by-row lighting mode. The FPGA clock is 50 MHz, and 320 ns is required for the circuit to scan all rows.