This paper gives the theory on intelligent optimization of column temp
eratures in a given gas chromatographic column system. Based on the st
ructure-retention relationship between solutes and solvents, and the d
ependence of retention values on temperatures, most difficult separati
ng component pairs and temperature separability range can be predicted
, ideal optimum temperature programming, in which every step is determ
ined by a most difficult separating component pair, can be further obt
ained.