The paper presents a new dual-mode nonlinear model predictive control (NMPC) scheme for constrained continuous-time nonlinear systems. The tool of control Lyapunov functions of nonlinear systems is used to compute the terminal constrained regions and terminal control laws with free-parameters within the dual-mode NMPC framework. By offline selecting the parameters in the terminal control law increases the size of the terminal region. Moreover, in order to obtain the optimality of the terminal control law with respect to given cost functions, the parameters are optimized online. Then a vary time-horizon dual-mode NMPC algorithm is formulate for the constrained nonlinear system, whose recursively feasibility and closed-loop stability properties are also derived. Finally, an example of the spring-cart control system is used to demonstrate the effectiveness of the resulted obtained here, where the linearization LQR method to be exploited to compute the terminal region of the spring-cart control system is compared and studied.