Introduction: Valproic acid (VPA), an antiepileptic major, is widely used in the generalized and partial epilepsy. The need for determination of plasma antiepileptic including the VPA was felt when it was found that the dose is not always predictive of therapeutic effect. Methods and Results: We collected 49 patients (mean age 29 ± 14.6 years), who have benefited from 51 dosages. The indication of VPA plasma monitoring was occurrence of side effects (45%), uncontrolled epilepsy (31.3%) and suspicion of poor adherence to treatment (7.8%). The VPA was prescribed as monotherapy (57%) or in combination (43%). The plasma concentration of VPA was in the therapeutic range in 64.5%, sub-therapeutic 21.5% and above 14% supra-therapeutic. Adverse events that have been observed in our patients were multiple. The most frequent were trembling (37%), cognitive disorders (25.5%) and obesity (19.5%). The risk of side effects was related mainly to a high plasma concentration. Indeed, 71.4% of patients who have a plasma supra-therapeutic VPA, have developed side effects. In our study 11 patients (21.5%) had a rate below therapeutic plasma. Three of these patients had very low plasma levels, probably reflecting a lack of dosing and the four others had a low rate reflecting an irregular dosing. Conclusions: The determination of plasma antiepileptic is not systematic; it is carried out when making a response for a clinician question, for having the best cost / benefit. The adjustment therapy should only occur if the need in the clinic was felt. Indeed, the VPA plasma monitoring avoids the toxic concentrations (overdose) or otherwise ineffective (under-dosing). This allows better control of seizures with minimal risk of side effects and lower cost. The interpretation of the results of assays should be prudent and flexible and therefore requires close collaboration between the clinician and pharmacologist.