Abstract: After the modified Q vector (Q*) is transferred and handled, a kind of Q vector (QN) is derived, whose calculating formula is similar to the counterpart of quasi-geostrophic Q vector (QG), however, QN vector is calculated using actual winds. The diagnostic abilities of QN vector and QG vector are compared based on a typical Changjiang-Huaihe Meiyu front cyclone (MYFC) rainstorm occurred during 5th-6th July 1991, and the results show that the diagnostic ability of QN vector is better in comparison to the counterpart of QG vector. In the meantime, the convergence field of QN vectors at 700 hPa can serve as a good indicator to the horizontally distributive character of the simultaneous actual precipitation. Then, the QN vector is partitioned into four components: QNalst (the “alongstream stretching” component that represents the intensifying/weakening of the alongflow thermal gradient by contraction/stretch of isotherm spacing), QNcurv (the “curvature” component that represents the curvature effect, according to which a downstream increase (decrease) in the cyclonic curvature of the isohypses induces subsidence (ascent)), QNshdv (the “shear advection” component that represents the thermal advection by horizontal wind shear), and QNcrst (the “crossstream stretching” that represents the effect of confluence and diffluence of the winds. i.e. the strengthening/weakening of the cross-flow thermal gradient forced by confluence and diffluence of the winds) in the natural coordinate system that follows the isohypses (hereafter referred to as PG partitioning). The application of QN PG partitioning to the analysis of the MYFC torrential rain from 2000 BST 5th to 2000 BST 6th 1991 indicates that the QN PG partitioning was able to disclose the potential physical mechanism of synoptic process, which was difficult for “total” QN (which is equal to QN) to display. In particular, the horizontally distributive characteristic of was always similar to that of , and accounted for a large portion of at the different stages of MYFC, which means that the former had large contribution to the latter in the context of exciting and forcing vertical motion generation and played a dominant forcing role in the formation of MYFC precipitation. During the whole evolution of the heavy rain of MYFC, the effect of QNcurv on the genesis of rain got little by little, until played a suppressing role. While the positive forcing effect of QNshdv on the genesis of precipitation got obviously stronger with the development of MYFC and got rapidly weakened, until almost disappeared as the MYFC moved eastwards to the sea and decayed. QNcrs had no effect on the generation of the precipitation during the developing and intensifying stages of MYFC but played a leading role at the decaying stage. In addition, the divergences of QNalst and QNcurv, and QNshdv and QNcrst showed similar horizontally distributive features, respectively, and no obviously mutual cancellation, but differences in intensity during the developing and intensifying stages of MYFC, however, at the decaying stage they did display the opposite distributive features and mutual cancellation. Obviously, the QN PG partitioning can disclose clearly different forcing factors of precipitation generation at different stages of MYFC.