Based on the distance between atoms and firstly electronegativity of each atom, a new set of descriptors called molecular electronegativity edge-distance vector (VMED) applied to describe molecular structure of E.coli transcriptional DNA-promoter sequences, is firstly proposed , because it is easy to calculate, only from primary structure of DNAs. Here a new type of quantitative structure-activity relationships (QSARs) called quantitative sequence-property models (QSAM) is developed with good forecasting ability by multiple linear regression (MLR) method. The results show that VMED has both excellent structural selectivity and good activity estimation. Besides, this novel vector will be useful to structure characterization and activity prediction of biological molecules. The resulting structural descriptors can be used to investigate requirments for new nucleic acids(NAS) in order to obtain sequences with altered activities.