Abstract: ObjectiveTo study the computing methods of the combined-variables in multivariate analysis of the quality of survival.MethodsThe nonlinear principal component analysis is used to composition variables measuring the quality of survival and linear regression is used to composition influencing variables on the survival quality.ResultsThe variance proportion drawn by the first principal component is 0.9948 and weights of three combined-variables composed by influencing factors on the survival quality are 0.35,0.32 and 0.33.ConclusionUsing nonlinear principal component analysis and linear regression to computing combined-variables are suitable to the influencing factors analysis on the survival quality.