Ab initio calculations at the second-order Moller-Plesset perturbation theory (MP2) level have been performed on the cis and trans isomers of the title model complex. For the cis isomer, the optimized structure shows a H...Si interaction distance of 1.813 Angstrom, in close agreement with the experimental value of 1.769 Angstrom for cis-Mo(CO)(depe)(2)(H...SiHPh(2)). Analysis of the density Laplacian (del(2) rho) around the Mo-(H...Si) triangle reveals a predominantly covalent Mo-Si bond and a likely dative H--->Mo bond, with a heavily curved H...Si bond path, suggesting that the H...Si covalent interaction is severely weakened. The trans structure is 9.9 kcal/mol less stable than the cis isomer, caused by the competition of metal d-electrons between CO and H...Si, and is manifested by the shorter H...Si distance of 1.715 Angstrom. The inability to locate a stable 7-coordinate structure with a H...Si distance greater than 2.5 Angstrom suggests that a classical-nonclassical tautomerism analogous to that in some dihydrogen complexes seems not very likely, and the optimized structure for a pentagonal bipyramidal 7-coordinate isomer with H and Si not neighboring is 6.5 kcal/mol less stable than the observed cis isomer. Study of the rotation of the H...Si unit around the Mo-(H...Si) "bond" gives a rotation barrier of about 10 kcal/mol caused by the competition of Mo d-electrons for back-bonding similar to that for the trans isomer. On the basis of molecular orbital arguments, both the Mo complex and the widely studied piano-stool complexes Mn(Cp)(CO)(2)(H...SiR(3)) can be regarded as pseudooctahedral d(6)-ML(6) which is a particularly stable class of complexes. The difference of the nature of bonding between H-H and H-Si complexes is also discussed in terms of orbital interaction.