Aiming at the imporvement of dynamic fuzzy neural network

a new algorithm is put forward

which combines the pruning strategy using column pivot SVD-QR method with the parameters adjustment. The pruning strategy with column pivot SVD-QR method is used to extract the most important fuzzy rules from a given rule base

which prevents the network structure from sustainable growth

and avoids the overfitting and over-training phenomenon. Extended Calman filtering method is utilized to divide the global algorithm into linear and nonlinear parts. The linear parameters and nonlinear ones can be updated respectively

thus the fast learning speed is achievable. The validity of the proposed algorithm is verified through the control of blood pressure. The results prove that the D-FNN with the pruning strategy using the column pivot SVD-QR method and the parameters adjustment has a good performance.