If you need to heparinize blood withdrawn from human simply take 0.5 mL of heparin into a 10 mL syringe, remove the syringe from heparin vial, pull the piston to 10 mL (aspirate air), agitate the syringe, push the piston and discharge all the heparin. The syringe needs only to be wet with heparin!

First, I would suggest that you consider using EDTA instead.  It became generally regarded as 'preferable' to use EDTA as the anticoagulant, especially if you don't know ahead of time what proteins you are interested in. Less potential for interferences with EDTA (1.8mg/mL whole blood).  For proteomics research, the selection of anticoagulant definitely matters, and it's not trivial to make the right choice!!  I have several publications regarding anticoagulant selection for proteomics if you are interested -- listed in my profile. I don't know anything about zerbrafish blood biochemistry, but if there are Ca++-dependent coagulation enzymes in the clotting cascade, EDTA would definitely be my choice.  (And given the conservation of so many Ca-dependent enzymes, the usual EDTA concentrations likely suffice, regardless of species.)  My previous work demonstrated the importance of including protease inhibitors in addition to EDTA.  As with all science, other researchers had other opinions (which anticoagulant, whether or not to use protease inhibitors).  Given that zebrafish are tiny, I doubt that you have the chance to collect and archive several different samples, with different anticoagulants, which would definitely be the safest approach (and meaning you would have that many more samples to study later too).

Heparin and proteomics:  If you intend to do mass spec, heparin can make a mess of the spectra, since it is an indeterminate-length polymer, and thus you will get TONS of peaks representing each incremental monomeric species.  This made a mess for some of my previous peptide analyses, where we did not do any LC first.  In this manner, it should is obvious that 'excess' heparin will further obscure mass spec profiles.  Since heparin is highly charged, it does get depleted easily on many LC schema.  For gels/DIGE, heparin shouldn't be as much of a problem per se, but do remember that it's high charge means that it may complicate gel loading, especially if you use large doses.

My experience with careful studies of heparin is only with human - which generally work just fine with around 10-15 Units of heparin per mL, per my recollection.  You could always contact a coagulation lab to check on coag parameters on your zebrafish samples, and get help on titrating to define what is 'enough' heparin for your purposes.

As a physiologist and a teacher of anatomy and physiology, I would go with EDTA as well. I collect on a regular basis blood from Xenopus, chicken and various mammals (not zebrafish though) and in these species, some coagulation will still occur with either Heparin or citrate. EDTA works a lot better.

I would agree with EDTA to anti-coagulate blood samples. It is also used as an anti-coagulant in tubes where blood is drawn for performing a full blood count in automatic analyzers ! EDTA is a small molecule that chelates Ca++ ions without interfering with blood cells or proteins in plasma.