Research on the low voltage power line communication (PLC) modulation recognition technology is of great importance for establishing PLC physical layer standard test method. In this paper, a novel modulation recognition scheme with double feature parameters-threshold judgment method for narrowband PLC is proposed. The PLC modulated signal is sampled and pre-processed. With the four decision criteria derived from these samples, the BP neural network is applied to recognize and classify different modulated signals automatically. From the experiment results it can be seen that the proposed method based on double feature parameters-threshold judgment can effec-tively recognize PLC narrowband signals. When the SNR is above 8dB, the successful recognition rate is above 95%.