There is no one best method for soil sampling in rice, or any other field. It is dependant on the area (area) of the land, topography, type (how detail) of analyses and such other considerations.
That should be part of the leaf harvest (agriculture)
Rice: highest leaves before flowering (http://gtalk.ir/archive/index.php/t-290444.html)
Enhancing Indigenous Knowledge in Rice Terraces: Recently, the threat of climate change looms ever larger and is widely projected to impact the frequency and intensity of extreme weather events. The expected increase in extreme rainfall and resultant water cycle changes are likely to test the existing capacity of rice terraces.
To get the soil samples, it has to considere the sampling intensity, time budget, and the purpose of the research. The most important thing is the representativeness of the samples. So, at the area of rice field terraces, the soil samples can be taken from the hill, middle and lower area, and at least please make 3 replication.
It would be more helpful if you could provide the following:
1- Soil properties you would like to analyze based on your objectives.
For example, if you were to do pH, EC, macro and micro nutrients, SAR, particle size, water holding capacity etc. you should know the quantity of soil required for each element and therefore collect the appropriate quantity in the field.
2- Type of experimental design (CRD, RCBD, GRCBD etc.). Is there a blocking factor or not? What are your treatments? What are your experimental and observational units?
3- The number of observational units (might be your plots) and their size: the number of samples should be representative to the total number of observational units. The size or shape of your plots should guide on the number of replicates you will need within each plot.
You should also consider your budget as Ambar has suggested. It is important too.