Blood (serum/plasma) antiepileptic drug (AED) therapeutic drug monitoring (TDM) has proven to be an invaluable surrogate marker for individualizing and optimizing the drug management of patients with epilepsy. Since 1989, there has been an exponential increase in AEDs with 23 currently licensed for clinical use, and recently, there has been renewed and extensive interest in the use of saliva as an alternative matrix for AED TDM. The advantages of saliva include the fact that for many AEDs it reflects the free (pharmacologically active) concentration in serum; it is readily sampled, can be sampled repetitively, and sampling is noninvasive; does not require the expertise of a phlebotomist; and is preferred by many patients, particularly children and the elderly. For each AED, this review summarizes the key pharmacokinetic characteristics relevant to the practice of TDM, discusses the use of other biological matrices with particular emphasis on saliva and the evidence that saliva concentration reflects those in serum. Also discussed are the indications for salivary AED TDM, the key factors to consider when saliva sampling is to be undertaken, and finally, a practical protocol is described so as to enable AED TDM to be applied optimally and effectively in the clinical setting. Overall, there is compelling evidence that salivary TDM can be usefully applied so as to optimize the treatment of epilepsy with carbamazepine, clobazam, ethosuximide, gabapentin, lacosamide, lamotrigine, levetiracetam, oxcarbazepine, phenobarbital, phenytoin, primidone, topiramate, and zonisamide. Salivary TDM of valproic acid is probably not helpful, whereas for clonazepam, eslicarbazepine acetate, felbamate, pregabalin, retigabine, rufinamide, stiripentol, tiagabine, and vigabatrin, the data are sparse or nonexistent.