Genetic information, which is mostly encoded in the form of DNA sequence, is the basis of life. Its deviations are often the cause of the most deadly diseases such as cancer. Accordingly, the development of methods to control the transcription of certain DNA parts is an important direction of modern pharmacological and biological research. Within the scope of this work, we are investigating the second generation of a polyintercalating agent that we developed earlier, potentially capable of recognizing 16-bp DNA sequences. In order to confirm its ability for advanced selective DNA recognition a series of simulation experiments was conducted. We differentially investigated the stability of HASDI-G2 complexes with mutated targeting sequences and their native variants. Firstly, we confirmed the ability of HASDI-G2 to clearly discriminate the target sequence (EBNA1) from a random site in the human genome (KCNH2). That repeated the experiment of the polyintercalator's previous version and additionally showed better results of the next-generation structure. Next, we examined HASDI-G2 under conditions where the target sequence differed from the random one increasingly slightly. And we found that even a one-nucleotide mismatch leads to a similar complex destabilization as a mismatch of 3 or 4 nucleotides. Such complexes showed significant conformational rearrangements, accompanied by a sharp decrease in the hydrogen bonds quantity, a drop in the binding free energy, and local melting of the DNA duplex. Moreover, the latter applied not only to sites of direct incompatibility, but also to parts where HASDI-G2 fully corresponded to the sequence of intercalation.