In order to overcome the shortcomings of slow convergence speed and large steady-state error of Constant Modulus Algorithm(CMA),on the basis of analysis of constant modulus blind equalization algorithm based on variable segment error function,a constant modulus blind equalization algorithm based upon dynamic variable segment error function(DVSCMA) was proposed.The blind equalization algorithm adjusts the section position of the error function dynamically by using the Mean Square Error (MSE),and the characteristics of the error function is changing in the balancing process with the MSE,which enables the error model and the transmit signal model to match continuously,thus accelerate the convergence speed and reduce the steady state error.The new algorithm is simulated by using mixed-phase and minimum phase of underwater acoustic channel.The simulation results demonstrate that the proposed algorithm has significantly faster convergence rate and lower steady-state mean square error compared with the CMA for the mixed-phase of underwater acoustic channel,while the new algorithm has significantly lower steady-state mean square error and very close convergence rate compared with the CMA for the minimum phase of underwater acoustic channel.