After a failure of "One-Size-Fits-All" approach on traditional treatment of diseases, the age of multi-faceted personalized medicine came into being for tailoring therapy to get the best response in the highest safety margin. Personalized drug dosage is one of the promises of personalized medicine and is of great importance for critical medical treatments such as anesthesia. General anesthesia induces a reversible state of unconsciousness in the patient due to the synergetic action of a cocktail of drugs. A general anesthesia cocktail is made up of a hypnotic, such as propofol, an analgesic, such as opioid-based or paracetamol, and a myorelaxant drug, such as midazolam. The delivery rate of these individual drugs in the cocktail has to be continuously monitored and balanced to achieve and maintain the desired level of sedation in the patient to avoid severe psycho-physical complications of over-/under-doses. To this end, the common practice nowadays is to use Target Control Infusion pumps, which regulate the delivery of anesthetic compounds according to their mathematical pharmaco-kinetic models that estimate the right dose for the injection on the basis of patients’ health data. However, implementation of these models have been done considering a population that does not cover all individuals, hence, falling in the category of "One-Size-Fits-All" approach and fail to reproduce the inter-patient variability in metabolism. Therefore, one of the main challenges in general anesthesia practices is the monitoring of the Depth Of Anesthesia in patients to be able to tune the drug dosage according to individual’s responses. One of the most widely used methods for DOA monitoring in Europe is the BiSpectral Index, a statistical predictor that is evaluated from ElectroEncephaloGram signal. However, there exist several limitations regarding this technique such as being indirect measure of DOA and prone to be affected by artifacts and production of different signals in case of different anesthetics. To overcome these limitations, there is a need for a Therapeutic Drug Monitoring system capable of continuously measuring the actual concentration of each infused drug in the blood of patient during DOA for a safer, reliable and personalized anesthesia delivery. In this thesis, we present the design, the implementation and the validation of a complete TDM system for long-time and continuous monitoring of anesthetics’ concentrations. The complete TDM system consists of: • a custom-built Raspberry-Pi-based electronic board to drive, • the electrochemical sensors for propofol, paracetamol and midazolam detection, that are encapsulated into • a fluidic device to drive the sample on the sensing sites, and • an Android-based Internet of Things network architecture to keep the anesthesiologist always connected with the sedated patient. Thanks to the IoT network, the developed system is capable of fast data visualization tools, as well as an alarm system activated in case serious physical conditions are detected in patient and remote monitoring through cloud-sharing. This thesis describes in details the system and highlights the following results:•Fouling-free propofol long-time monitoring by Pencil Graphite Electrode with specific lead composition. •Assembly of electronics and sensors in an innovative IoT multi-panel system for monitoring several drugs in complex bio-fluids. •Demonstrate the successful monitoring of several drug concentrations over time