ABSTRACT The luminosity function of Fast Radio Bursts (FRBs), defined as the event rate per unit cosmic co-moving volume per unit luminosity, may help to reveal the possible origins of FRBs and design the optimal searching strategy. With the Bayesian modelling, we measure the FRB luminosity function using 46 known FRBs. Our Bayesian framework self-consistently models the selection effects, including the survey sensitivity, the telescope beam response, and the electron distributions from Milky Way/ the host galaxy/ local environment of FRBs. Different from the previous companion paper, we pay attention to the FRB event rate density and model the event counts of FRB surveys based on the Poisson statistics. Assuming a Schechter luminosity function form, we infer (at the 95 per cent confidence level) that the characteristic FRB event rate density at the upper cut-off luminosity $L^*=2.9_{-1.7}^{+11.9}\times 10^{44}\, \mathrm{erg}\, \mathrm{s}^{-1}$ is $ϕ ^*=339_{-313}^{+1074}\, \mathrm{Gpc}^{-3}\, \mathrm{yr}^{-1}$, the power-law index is $α =-1.79_{-0.35}^{+0.31}$, and the lower cut-off luminosity is $L_0\le 9.1\times 10^{41}\, \mathrm{erg}\, \mathrm{s}^{-1}$. The event rate density of FRBs is found to be $3.5_{-2.4}^{+5.7}\times 10^4\, \mathrm{Gpc}^{-3}\, \mathrm{yr}^{-1}$ above $10^{42}\, \mathrm{erg}\, \mathrm{s}^{-1}$, $5.0_{-2.3}^{+3.2}\times 10^3\, \mathrm{Gpc}^{-3}\, \mathrm{yr}^{-1}$ above $10^{43}\, \mathrm{erg}\, \mathrm{s}^{-1}$ , and $3.7_{-2.0}^{+3.5}\times 10^2\, \mathrm{Gpc}^{-3}\, \mathrm{yr}^{-1}$ above $10^{44}\, \mathrm{erg}\, \mathrm{s}^{-1}$. As a result, we find that, for searches conducted at 1.4 GHz, the optimal diameter of single-dish radio telescopes to detect FRBs is 30–40 m. The possible astrophysical implications of the measured event rate density are also discussed in the current paper.