The effects of reverse reactivity stratification, i.e., port-injection (PI) of n-heptane with direct-injection (DI) of iso-octane and PI of iso-octane with DI of n-heptane, on combustion characteristics and flame development were investigated. The volume fractions of port-injected fuel were 30%, 70% and 90%. The DI timings were −25°, −15° and −5° CA ATDC. The optical diagnostic methods of high-speed imaging and flame emission spectra were adopted. Results show when n-heptane is port injected and iso-octane is direct injected, the PI volume fraction plays more important role in controlling combustion phasing, while the DI timing is the dominant factor on flame development structure that can be classified into four classes. When iso-octane is port injected and n-heptane is direct injected, the DI timing dominates the combustion phasing. Flame kernels merge to form larger flame area with reactivity stratification introduction compared with the condition that nearly only one initial flame kernel forms in each downstream spray without reactivity stratification introduction. The initial mean flame speed is around 80 m/s with reactivity stratification introduction, which is nearly twice faster than that without reactivity stratification introduction. Finally, when comparing conditions of reverse reactivity stratification, it provides guidance for dual-fuel combustion engines that the overall reactivity cannot be regarded as the solely important parameter to optimize combustion phasing and other combustion characteristics. At some extent, the effects of overall reactivity on combustion characteristics are lower than the effects of volume fraction, equivalence ratio, injection order or start of injection (SOI) timing of high reactivity fuel.