There is a nice paper that has been recently published by Chaudhry ZZ et al. in Lab. Anim. (Oct 2013). http://www.ncbi.nlm.nih.gov/pubmed/23760565. Fasting is usually used to minimize competition between STZ and glucose for low affinity GLUT2 transporters on β cells.

Other nice paper with a specific topic about Streptozotocin administration at fasting state is "Review of the mechanism of cell death resulting from streptozotocin challenge in experimental animals, its practical use and potential risk to humans" published by Eleazu et al.

http://www.ncbi.nlm.nih.gov/pubmed/24364898

Fasting is not necessary to have rats become diabetic by injecting STZ.

However, to get consistent and reliable results, many factors should be considered and they all relate to available dose of STZ.

Here are a few things to consider (I am assuming a single STZ treatment):

1. Route of delivery (intravenous versus intraperitoneal- if you are proficient in iv injections, results are more consistent; if not, consider ip)

2. Fasting versus not fasting, plays into the matter of consistency of results - with fasted animals, if a diabetogenic STZ dose is delivered correctly (freshly prepared solution injected within a s few minutes of dissolution), you will always have animals become diabetic.

If you are trying to use STZ treatment as a model of type 1 diabetes, I am not sure what to make of the paper Ercument Dirice linked where the zero time blood glucose level of normal control mice in GTT is around 200mg/dl, all so-called diabetic mice at d10 are also the same (~200mg/dl) and remaining beta cell area is over half of the normal controls.