The contribution deals with the bayesian point estimation of the Poissons´ parameter, for which the conjugate family Poisson/gamma is usually used in bayesian statistics. Presented aaproach is based on this conjugate family, too, but differs in the criterion of point estimations´quality - in the contribution the mean square error is used.Detailed analysis of relations among the variables occured in model allowed to develope the algorithm for determination the bayesian point estimation. As a great benefit the possibility of simple definition of prior gamma distributions´parameters appears. The consequence is such bayesian point estimation, the mean square error of which is smaller than that of the classical point estimation within the pre-determined interval, which corresponds with the prior conception of estimated parameter.
