In situ organic matters, such as Mucuna sp., Flemingia sp., and plant residues differ in quality (in terms of lignin, cellulose and nutrients contents). Such difference determines soil organic matter content especially labile fraction that could effect on soil properties. The aims of this research were to study: 1) the effects of various sources and quality of organic matters on soil organic fraction, and 2) the relationships between the changes of particulate organic matter and C-microbes (POM and Cmic) and soil quality indicators. The experiments were conducted in greenhouse of the Center for Soil and Agroclimate Research and Development, Bogor, from January to May 2003. The factorial completely randomized design was used with three replications. The first factor was the method of organic matters application: 1) mulch on soil surface and 2) incorporated with soils. The second factor was the sources of organic matters (dosage in equvalent with 2% organic C), i.e: 1) Mucuna sp., 2) Flemingia sp., 3)maize residue, 4) mixture of Mucuna sp., and Flemingia sp., 5) mixture of Flemingia sp. and maize residue, 6) mixture of Mucuna sp. and maize residue, and 7) mixture of Mucuna sp., Flemingia sp., and maize residue. The results showed that on organic matter application with incorporated method, C/N ratio and lignin content of organic matter have positive correlation with organic C, Cmic and Cmic/Corg, but have negative correlation with POMp. The changes of soil qualities caused by short term organic matter application could be seen from the changes of labile fraction of organic matter. When organic matter was incorporated, POMtand POMt/Corg have positive correlation with bulk density, but have negative correlation with RPT, PDC, available K, and total- N. When organic matter was spreaded on the soil surface, Cmic has positive correlation with available P. The changes on labile fraction (Cmic and POMt) was easy to be detected in short period, thus the changes of physical characteristics and the availability of macro nutrients is easy to be predicted.