The best to start with is the FAO suitability approach. The suitability is based on soil mapping units. You need this map with all soil properties then after that you need to see the units which are highly suitable. Of course you need climatic data for temperature and rainfall. This is the biophysical approach for fruit trees. I advise to consider socioeconomic factors to remove some constraints and limitations, e.g. the use of terraces to overcome slope, fertilization to overcome deficiency of some nutrients. I am attaching two articles as examples
Fao Suitability analysis FAO approach is definitely an initial step but according to Davidson, 1992; Rossiter, 1990, 'The FAO (1976) framework for land suitability involves the construction of matching tables or transfer functions, and subsequent calculations of suitability. However, these processes are liable to errors. Therefore, there are a great number of benefits to be gained in automating the FAO procedures'.
You should try the multicriteria evaluation using IDRISI software. It's a geographic supervised analysis to establish land cover suitability used for spatial location of predicted time transitions.