We have determined the Optical Luminosity Function (OLF) of a sample of 80946 Quasi Stellar Objects (QSOs) taken from the Sloan Digital Sky Survey Data Release Seven (SDSS DR7) with redshift range  0.3 < z < 2.5 and absolute magnitude M_i < -22.5. The Monte Carlo Technique of numerical integration is used. The sample of QSOs is divided into seven sub-samples with redshift in the ranges: 0.30 < z < 0.55, 0.55 < z < 0.80, 0.8055 < z < 1.05, 1.05 < z < 1.30, 1.30 < z < 1.55 , 1.55 < z < 1.80, and 1.80 < z < 2.05. Each redshift interval is binned in absolute magnitude with bin width ΔM_i = -0.5. A flat universe with cosmological parameters Ω_m = 0.3, Ω_∧ = 0.7, and Hubble constant Ho = 70.0 km·s^-1·Mpc^-1 is used. From the optical luminosity function a clear evidence of AGN downsizing is observed, i.e. the number density of the less luminous AGNs peaks at lower redshift than the number density of the more luminous AGNs.