Communication object recognition is a critical section in military operations. With the continuous improvement of counter reconnaissance technologies, communication circumstances are becoming more and more complex. Complicated and variable of the communication parameters, error or deficiency of reconnaissance data make it difficult for object recognition. An improved template-matching method to address the communication parameters variability and deficiency problems is presented. The simulation results show that this method has high recognition rates and false-tolerance capabilities, and can be used to increase object recognition rates under large parameters error and deficiency conditions.