Detailed mechanisms describing complex phenomena of combustion chemistry, such as flame propagation or pollutant formation, involve hundreds of species and thousands of elementary reactions and cannot be handled in practical simulations of turbulent combustion. A widely used way to reduce chemistry is to build look-up tables where chemical parameters such as reaction rates and/or species mass fractions are determined from a reduced set of coordinates (ILDM, FPI, or FGM methods). Nevertheless, these tables may require large memory spaces and nonnegligible access times, especially when running on massively parallel computers. In this work, the self-similarity behavior of laminar premixed flames is first put into evidence and then theoretically sustained. This property provides a way to reduce the size of chemical databases, especially for computations on massively parallel machines, under the FPI (flame prolongation of ILDM) framework. The database is reduced to similarity profiles for the species reaction rates (or the species mass fractions), stored together with scaling rules. This new formulation is then implemented in the PREMIX code and numerical simulations of laminar premixed flames successfully compare with full chemistry computation, validating this promising approach.