We perform numerical simulations of a seismic cycle by applying a laboratory-derived friction law to evaluate possible crustal deformations prior to the hypothesized Tokai earthquake, which is expected to occur in the near future along the Suruga trough, central Japan. The frictional force obeying a rate- and state-dependent friction law is assumed to act on the plate interface in a 2D model of the subduction zone. We test two versions of the friction law: the so-called slowness version and the slip version. We select model parameters on the basis of historical documents about past large earthquakes in the Tokai district and the present coupling state of the plate interface beneath the Tokai district estimated by GPS observations. Simulation results for both the slowness and the slip version are as follows: (1) The coupling region of the plate interface becomes narrower with time in the interseismic period. (2) The inland crust contracts at a nearly constant rate for most of the interseismic period. The contraction rate decreases or a change from contraction to extension takes place before the occurrence of a large interplate earthquake. (3) Significant abnormal crustal deformation precedes the earthquake, and maximum amplitudes about one day before the earthquake exceed the noise levels in the case of clear weather of borehole strainmeters installed in the Tokai district. (4) The spatiotemporal variation of aseismic sliding on the plate interface perturbs the regional stress field around the source volume of the coming earthquake, leading possibly to precursory seismic quiescence in the overriding plate and a change in focal mechanisms of small earthquakes in the subducting slab. The most significant difference in the results between the slowness version and the slip version of the friction law is that amplitudes of preslip and crustal deformation just before earthquakes are several times larger for the slowness version than for the slip version. Although some ambiguity remains in the present model mainly because of the incompleteness of the friction law and the deficiency of observational data constraining model parameters, the physical model can be improved through the comparison of predicted crustal deformation with observations. Continuous and concentrated observations of crustal deformation and seismic activity are thus important in the Tokai district to understand quantitatively the sliding process preceding a large interplate earthquake there.