A method of using asynchronous delay sampling and generalized regression neural network (GRNN) to monitor optical signal noise ratio (OSNR) and chromatic dispersion (CD) is proposed. In the 40 Gbit/s 16QAM communication systems and the nonlinear channel environment, the feature quantities are extracted by asynchronous sampling, and GRNN is used to realize OSNR and CD monitoring. Comparing with other neural network methods, GRNN method can not only realize the monitoring of OSNR and CD, but also has the advantages of fewer parameters, high efficiency algorithm and easy optimization.