The oxidative addition of an Si-X sigma-bond (X = H or Si) to M(PH3)(2) (M = Pd or Pt) is investigated with ab initio MO/MP4, SD-CI, and CCD methods. Geometries of reactants, transition states (TS), and products are optimized at the MP2 level. Although addition of an f-polarization function on Pt and Pd changes the activation energy (E(a)) and the reaction energy (Delta E) little, addition of a d-polarization function on P decreases the exothermicity considerably. Although the MP2-optimized geometries differ somewhat from the SCF-optimized ones, E(a) and Delta E calculated for the SCF geometries are almost the same as E(a) and Delta E for the MP2 geometries. E(a) and Delta E hardly change upon going from MP4SDQ to SD-CI and CCD, whereas they fluctuate somewhat at the MP3 and MP4DQ levels. The instability of Hartree-Fock wave function is not observed even at the TS. These results indicate that a single reference wave function can be used for these oxidative addition reactions. Unexpectedly, the Si-Si and Si-H oxidative additions to Pd(PH3)(2) proceed with a lower E(a), than those to Pt(PH3)(2), while the former are less exothermic than the latter. These results are explained in terms of bond energies and the electronic structure of the TS.