Objective In order to solve the problems of the difficulty and low efficiency in the recognition of one-dimensional barcodes under tilt conditions, a recognition method using the matching translation was proposed. Methods Firstly, the image of barcode underwent conventional preprocessing and run-length encoding, and then two parallel matched lines were selected. The intersections of the two parallel scanning lines and the barcode strip or empty crosspoint were used to calculate the slope of the barcode, then the best barcode scanning line could be achieved using the slope. Finally, the barcode could be scanned again and successfully decoded. Tests of sample barcodes were conducted. In the environment of C++ Builder 6, the barcode recognition algorithm was realized. Results The experiments proved that the method could effectively identify the barcodes under any tilt conditions and the correct recognition rate could reach higher than 98% . Conclusion Compared to the traditional method, the calculation amount was obviously reduced, the recognition speed was significantly improved, and the recognition time was less than 5 ms generally.