Developed RBFNN and GRNN models with higher accuracy and quicker convergence when compared with other supervised learning network.

The normalized acoustic emission (AE) cumulative parameters are considered for developing the models.

The input AE parameters are considered only in the constant region (500–800 ms) of every 1% of ultimate load (40–70%) for minimising an error.

The mean and median of predicted failure load is also computed and compared with experimental failure load.

Degradation behavior of glass/epoxy composite laminates exposed to seawater environment has hardly been investigated.