Using a natural abundance C-13 method, soil organic matter (SOM) decomposition was studied in a C-3 plant - 'C-4 soil' (C-3 plant grown in a soil obtained from a grassland dominated by C-4 grasses) system and a C-4 plant - 'C-3 soil' (C-4 plant grown in a soil taken from a pasture dominated by C-3 grasses) system. In C-3 plant - 'C-4 soil' system, cumulative soil-derived CO2-C were higher in the soils planted with soybean (5499 mg pot(-1)) and sunflower (4484 mg pot(-1)) than that in 'C-4 soil' control (3237 mg pot(-1)) without plants. In other words, the decomposition of SOM in soils planted with soybean and sunflower were 69.9% and 38.5% faster than 'C-4 soil' control. In C-4 plant - 'C-3 soil' system, there was an overall negative priming effect of live roots an the decomposition of SOM. The cumulative soil-derived CO2-C were lower in the soils planted with sorghum (2308 mg pot(-1)) and amaranthus (2413 mg pot(-1)) than that in 'C-3 soil' control (2541 mg pot(-1)). The decomposition of SOM in soils planted with sorghum and amaranthus were 9.2% and 5.1% slower than 'C-3 soil' control. Our results also showed that rhizosphere priming effects on SOM decomposition were positive at all developmental stages in C-3 plant - 'C-4 soil' system, but the direction of the rhizosphere priming effect changed at different developmental stages in the C-4 plant - 'C-3 soil' system. Implications of rhizosphere priming effects on SOM decomposition were discussed.