why band gap reduces when pressure increased for provskite compounds ?
It should always be a practice to provide exact citation to a published paper when you're asking something of this nature.
Without knowing exactly the details, and at what level you're asking this, the generic Hubbard model for the band structure contains a U/W ratio, where U is the on-site Coulomb repulsion, and W is the bandwidth. When you increase pressure, the ions in the lattice are pushed closer, and thus, it increases U. If the bandwidth doesn't change, then the ratio U/W also increases. However, this usually translates into an increase in the size of the bandgap, i.e. the bottom of the conduction band moves away from the top of the valence band.
The bandwidth can easily change with pressure, so the ratio of U/W may behave differently for different compounds. But without proper citation, I do not have an explanation for your question.