A problem of regulatory single nucleotide polymorphism (SNP) prediction based on the presence of exomic SNPs is considered. Usually, regulatory polymorphisms are located in the non-coding region of DNA, therefore they cannot be determined during RNA sequencing. Thus, there is a problem of regulatory polymorphisms prediction by the presence of certain polymorphisms in the DNA coding region. The prediction can be made if there is a strong relationship between regulatory and exomic polymorphism. Chi-square test statistic and normalized Hamming distance for linkage disequilibrium are compared. The results of study of relationship of some regulatory SNP with exomic SNP are presented. To study the relationship between polymorphisms, the polymorphisms already found from the project database of 1000 genomes have been used. The usage of regression models for this problem is discussed. In general, there is strong relationship between regulatory SNP and exomic SNP and linear regression model significantly describes the relationship between regulatory SNP and exomic SNPs very well, however in some cases usage of more complicated types of regression models is necessary.