I do it through a set of scripts executed by a script which is configured through config files. Schema changes made according to the change log, code is redeployed at every update completely. All the changes/code are introduced in form of script registered in one of the config files. All the code is under source control - subversion on my machine and/or TFS for the projects. Most of the time I don't even know were my databases are deployed - developers and testers are able to install and update their local and shared copies by themselves. I described it in an article here: http://www.simple-talk.com/sql/database-administration/deploying-database-developments/[^] Please mind that while the article decribes approach that I'm using allover the place for a few years, the script attached to the article was just rewritten and was pretty buggy :). I have an updated version of it, should you be interested in it - email me. But the question is may be too broad - I use Erwin for the model, generate schema of it, put it to the scripts. For the changes usually use red-gate sql compare and modify it as needed. For the procs I use Management studio + sql toolbelt + CLCL utility (for the whole wack of templates for any occasions). Security and other config files are xml files, so I edit it with something. HTH