Sandy loess has high sand content and loose porous structure, and its mechanical properties are different from typical silty loess and sand. In order to study the effect of calcium source and curing method on the curing effect of MICP ( Microbially Induced Carbonate Precipitation ) technology in the treatment of sandy loess, the small-scale test was carried out by soaking method and perfusion method with three calcium sources of calcium chloride, calcium acetate and calcium lactate, respectively. The calcium carbonate content, calcium carbonate distribution uniformity and unconfined compressive strength of the solidified samples were used as evaluation indexes for comparative analysis. The results show that the uniformity of solidified samples is better when calcium acetate is used as calcium source. The calcium carbonate content and unconfined compressive strength of the solidified samples were higher when calcium lactate was used as the calcium source in the early stage of treatment, but with the increase of treatment time, the strength of the solidified samples with calcium chloride as the calcium source was higher. Compared with the soaking method, the perfusion method can effectively improve the poor seepage of calcium lactate, and the curing effect is better than that of calcium chloride and calcium acetate when calcium lactate is used as the calcium source. The calcium carbonate distribution of the solidified sample by perfusion method is more uniform and the strength is higher.