Sixteen indole derivatives have been computed at B3LYP/6-311G** level using density functional theory (DFT). Based on linear solvation energy theory, the structural para- meters were employed to present correlation between the parameters of chromatograph capacity factor (CCF) and molecular structural parameters. As a result, the correlation equation of the reversed phased high performance liquid chromatograph capacity factor to the intercept lgk’w and slope S of CCF were obtained, from which the correlation coefficients of lgk’w to the structural parameters are r2 = 0.9596 and q2 = 0.9262. While the correlation coefficients of the parameter S with structures are r2 = 0.9750 and q2 = 0.9252. Moreover, the effect of water as solvent on the present two models was also considered using SCRF method, and the result shows that the predicting capacity of correlation equation of lgkw’ increases, while that of the model for S decreases slightly. Both two correlation equations achieved in this work are more advantageous than those using theoretical descriptors from molecular connectivity indices. Sixteen indole derivatives have been computed at B3LYP / 6-311G ** level using density functional theory (DFT). Based on linear solvation energy theory, the structural para-meter were employed to present correlation between the parameters of chromatograph capacity factor (CCF) and molecular structural parameters. As a result, the correlation equation of the reversed phased high performance liquid chromatograph capacity factor to the intercept lgk’w and slope S of CCF were obtained, from which the correlation coefficients of lgk’w to the structural parameters are r2 = 0.9596 and q2 = 0.9262. While the correlation coefficients of the parameter S with structures are r2 = 0.9750 and q2 = 0.9252. Moreover, the effect of water as solvent on the present two models was also considered using SCRF method, and the result shows that the predicting capacity of correlation equation of lgkw ’increases, while that of the model for S decreases slightly. Both two correlation equations achieved in this wo rk are more advantageous than those using theoretical descriptors from molecular connectivity indices.