Question about model validation.

I'm working on a problem on my own (not for class or work) and I just wanted to get some outside opinions before I progress further. As far as my background, I've taken some introductory statistics courses for psychology majors, and an intro probability and statistics course for engineers. Additionally, I majored in physics, so I have a good math background. Overall, I just want advice on what statistical test I should use to verify that my model does a good job at predicting certain parameters. My model uses WinBUGS to perform a hierarchical Bayesian parameter estimation (so it's an MCMC). The model itself (which I won't go into here) is pretty complicated with a large number of parameters. In all likelihood, I'm going to have to estimate the parameters using what I've been told is bootstrap method (find one of the parameters, hold it constant; find another, hold both constant; find a third, hold that constant, etc). Therefore, I need to be able to say that at each step, the parameters are estimated accurately.

So far, my approach to testing whether the model is accurate is to
1) Generate the parameters (all parameters are continuous, not discrete) that I plan to estimate.
2) Create data from these parameters that would be data collected in an experiment.
3) Plug the data into WinBUGS. This returns the mean, median, standard deviation, and MC error.

I was thinking of doing a chi-square test (observed-estimated mean)^2/(standard deviation)^2 and summing over all the parameters of the same type. Before I waste a lot of time doing this (the model takes a while to run), I wanted to make sure that I'm verifying my model correctly. Is a chi-square the way to go or is there another test I should be using?