The Bellman function approach allows to prove many estimates in classical harmonic analysis Such as, for example, inverse H\"{o}lder inequality, Carleson imbedding theorem, Hunt-Muckenhoupt-Wheeden theorem and many many others. Moreover, this approach allows to get the correct dependence on parameters. For example, in such weighted problems as Hunt-Muckenhoupt-Wheeden theorem (and in related results) one can give the correct dependence of the estimate under consideration on the A_2 "norm" of the weight. Bellman function technique was invented by Burkholder to find the BEST constants in certain inequalities coming from martingale theory. But we will show in this course that the area of application of this technique is considerably wider than this.