A field experiment was carried out in 2005~2008 to a study effects of different tillage methods on soil microbial biomass C, N and P in plough layer of an oat field in Qingshuihe County, Inner Mongolia. The experiment was designed to have five treatments, i.e. non-tillage with short stubbles（NL）, non-tillage with tall stubbles (NH), non-tillage with short stubbles and mulch (NLS), non-tillage with tall stubbles and mulch (NHS) and conventional tillage (T). Results show no much difference between the treatments in trend in interannual variation of soil microbial biomass C, N and P in amount. Both soil microbial biomass C and N followed a double-peak curve. The peaks appreared at the jointing and filling stages, separately in Treatments NLS and NHS, and at booting and filling stages, separately in Treatments NL, NH and T while the peaks of soil microbial biomass N did at the seedling and filling stages, separately. However, the amount of soil microbial biomass P varied following a single-peak curve, with the peak appearing at the filling stage in all the treatments. Regardless of years or growing stages, the treatments followed a decreasing order of Treatment NHS > Treatment NLS > Treatment NH > Treatment NL > Treatment T in amount of soil microbial biomass C, N and P. All the non-tillage treatments affected yield of the crop, which declined first and then rose as compared with CK or Treatment T. Taking 2008 as an example, yield of the crop increased by 22%, 17%, 11% and 5%, respectively, in Treatment NHS, Treatment NLS, Treatment NH and Treatment NL. To sum up, non-tillage especially non-tillage with stubbles and mulch, is conducive to improvement of contents of soil microbial biomass C, N and P, and yield of the crop.