We present a new algorithm for isothermal-isobaric molecular-dynamics simulation. The method uses an extended Hamiltonian with an Andersen piston combined with the Nose-Poincare thermostat, recently developed by Bond, Leimkuhler, and Laird [J. Comp. Phys. 151, 114 (1999)]. This Nose-Poincare-Andersen (NPA) formulation has advantages over the Nose-Hoover-Andersen approach in that the NPA is Hamiltonian and can take advantage of symplectic integration schemes, which lead to enhanced stability for long-time simulations. The equations of motion are integrated using a generalized leapfrog algorithm (GLA) and the method is easy to implement, symplectic, explicit, and time reversible. To demonstrate the superior stability of the method we show results for test simulations using a model for aluminum and compare it to a recently developed time-reversible algorithm for Nose-Hoover-Anderson. In addition, an extension of the NPA to multiple time steps is outlined and a symplectic and time-reversible integration algorithm, based on the GLA, is given. (C) 2000 American Institute of Physics. [S0021-9606(00)51307-5].