We study the dynamical stability of planetary systems consisting of one hypothetical terrestrial-mass planet (1 or 10M_{earth}_) and one massive planet (10M{earth}-10M_jup_). We consider masses and orbits that cover the range of observed planetary system architectures (including non-zero initial eccentricities), determine the stability limit through N-body simulations, and compare it to the analytic Hill stability boundary. We show that for given masses and orbits of a two-planet system, a single parameter, which can be calculated analytically, describes the Lagrange stability boundary (no ejections or exchanges) but diverges significantly from the Hill stability boundary. However, we do find that the actual boundary is fractal, and therefore we also identify a second parameter which demarcates the transition from stable to unstable evolution. We show the portions of the habitable zones (HZs) of {rho} CrB, HD 164922, GJ 674, and HD 7924 that can support a terrestrial planet. These analyses clarify the stability boundaries in exoplanetary systems and demonstrate that, for most exoplanetary systems, numerical simulations of the stability of potentially habitable planets are only necessary over a narrow region of the parameter space. Finally, we also identify and provide a catalog of known systems that can host terrestrial planets in their HZs.