In the reifit process of a compressed natural gas (CNG) vehicle, the nozzle aperture of a CNG injector is an important process parameter and is difficult to determine. To improve the efficiency and accuracy of installation work, it is imperative to find a method for installers to quickly and accurately determine the required aperture. For this purpose, we analyzed a large volume of data obtained in actual refitting practices, and proposed a fuzzy pattern-recognition algorithm for determining the size of a nozzle aperture based on the engine power, the total mileage and the current mileage of spark plugs being used of the vehicle to be refitted. Verification of the algorithm in MATLAB finds that by using more random sampling and grouping and modifying the computing process of the algorithm, the accuracy of aperture recognition is improved significantly. The feasibility and accuracy of the algorithm are confirmed by comparison of the computed results with previous installation records.