From your question, it is not clear whether you want to:
1-avoid sending messages to malformed email addresses; or
2-avoid sending messages to email addresses which are not verified under your AWS account.
The answer for 1 is spread in different forms accross forums, SO, etc. You either do it simple, i.e., craft a short and clear regular expression which validates roughly 80% of the cases, or you use a very complex regular expression (in order to validate against the full compliance -- good luck, check this example), check whether the domain is not only valid but also up and running, and, last but not least, check if the account is valid under that domain. Up to you. I'd go with a simple regex.
The answer for 2 is available at Verifying Email Addresses in Amazon SES -- the Amazon SES API and SDKs support the operations below, so you should be covered in any case:
Using the Amazon SES API
You can also manage verified email addresses with the Amazon SES API. The following actions are available:
VerifyEmailIdentity
ListIdentities
DeleteIdentity
GetIdentityVerificationAttributes
Note
The API actions above are preferable to the following older API actions, which are deprecated as of the May 15, 2012 release of Domain Verification.
VerifyEmailAddress
ListVerifiedEmailAddresses
DeleteVerifiedEmailAddress
You can use these API actions to write a customized front-end application for email address verification. For a complete description of the API actions related to email verification, go to the Amazon Simple Email Service API Reference.