Using a statistical mechanical model and numerical simulations, we provide the design principle for the bridging strength (ξ) and linker density (ρ) dependent superselectivity in linker-mediated multivalent nanoparticle adsorption. When the bridges are insufficient, the formation of multiple bridges leads to both ξ- and ρ-dependent superselectivity. When the bridges are excessive, the system becomes insensitive to bridging strength due to entropy-induced self-saturation and shows a superselective desorption with respect to the linker density. Counterintuitively, lower linker density or stronger bridging strength enhances the superselectivity. These findings help the understanding of relevant biological processes and open up opportunities for applications in biosensing, drug delivery, and programmable self-assembly.