An organizational coevolutionary classification algorithm was proposed for recognition based on 1-D images of radar targets. It is different from available EA-based techniques mainly in that its evolutionary operations are performed on the examples directly, but not on the rules. It uses a bottom-up search mechanism, that is, it makes groups of examples evolved, and then rules are extracted from these groups of examples at the end of evolution. This method can avoid generating meaningless rules during the evolutionary process. The proposed method needs not extract the features and reduces the dimensions for high dimensional data previously. It has not complicated computations, with high training and recognition speeds result. Experimental results on the data of 3 airplanes obtained in a microwave anechoic chamber show that the proposed method has a stable performance and outperforms the methods based on SVMs and wavelet kernels. Its predictive accuracy is higher than 96%. In addition, its ability in resisting the noise is also tested, and a good result is obtained.