在车牌字符识别中，针对单一识别方法识别率不高的问题，提出了应用数据融合技术，将不同的识别方法有机地结合起来构成融合型识别系统，有效地、综合地提高整个系统的识别性能。数据层选择了加权平均算法，特征层选择了人工神经网络算法，决策层采用了模糊推理算法实现对车牌字符的最终识别。应用MATLAB进行了仿真，并与单独使用BP神经网络算法的识别率进行了比较，结果证明采用数据融合技术系统的识别率得到了较大提高，达到90%以上。 Abstract: According to the recognition rate of single recognition method is inefficiency in license recognition system，a new confluent license recognition method with data fusion technology was proposed to enhance the recognition rate effectively. In this method the weighted average algorithm，artificial neural network algorithm and fuzzy reasoning algorithm were used in data layer，feature layer and decision layer individually. At last，the system was compared with the BP neural network algorithm by MATLAB. The results show that the recognition rate has been greatly enhanced after adopting data fusion technology，and the recognition rate almost can reach to 90%.