Thanks for the answers. Global data might not be correct as my study area is located in Cambodia where the availability of data is less. I think I may need to do bias correction but I don't have observed data. If I don't correct the global data, how much percentage of reliable I may have?
There are meteorological drought indices, developed based on precipitation and evaporation. I wonder if these could be used in combination with soil characteristics to formulize certain types of agricultural drought indices for use in monitoring and studies of agricultural drought.
You can estimate soil moisture by using any soil water balance model.Thornthwaite and Mather model is a climatic water balance model and it will calculate soil moisture without considering the crop characters,It may not be a good model for assessing agricultural drought..FAO's water balance model may be intermediate in assessing agricultural drought. However, crop water balance models like RITCHIE model ,Aqua crop model,EPIC model etc will help in assessing agricultural drought.
You may want to use field-scale agricultural model to simulate soil moisture considering crop/plant growth, such as EPIC, APEX (field-scale version), GLEAMS, etc.
Sorry but the formulation of your question is definetively too much generic and, lets me say, a bit simplistic. You should better specify the geographical scale, the required resolution, the period of analysis, the level of accuracy, the soil layers and dephts of analysis, etc.
As is the question proposed is very complex provide you an exhaustive and effective answer.