For thousands of years music has accompanied human existence and development. Over the time it turned into a form of art capable of expressing beauty, ideals and emotions. Our previous work has investigated the possibility of automatic generation of music that would be (to some extent) alike to the music created by human composers. Our focus was on romantic era music, in particular a “Chopin-style” compositions. We have proposed a specifically designed memetic algorithm, which operated based on a handful of parameters and rules, which need to be followed when it comes to classical music. In this paper we review the proposed approach and extend it by introducing “a subjective factor” to the system, in the form of a specifically designed neural network. The role of this component is to provide the subjective preferences of the listener, which are taken into account during the music composition process. Preliminary results of this new system are presented.