A critical study of the different variables affecting the preparation of enzyme electrochemical biosensors using self-assembled monolayer (SAM)-modified electrodes is reported. Regarding variables affecting the obtention of SAMs on a gold disk electrode (AuE), the type of alkanethiol employed, its concentration, the time and temperature of formation and the composition of the medium were evaluated. Concerning a glucose biosensor based on SAM-modified AuEs, the performances of several redox mediators for the enzymatic oxidation of glucose were compared. Two glucose oxidase (GOx) immobilization methods on the SAM-modified AuE were tested: a covalent binding using 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide and N-hydroxysulfosuccinimide (NHSS), and a cross-linking with glutaraldehyde. The cross-linking method produced slopes of the substrate calibration graphs at a mercaptopropionic acid (MPA)-modified electrode two orders of magnitude higher than those obtained with the covalent binding. Moreover, the mediator tetrathiafulvalene (TTF) was co-immobilized atop the SAM together with GOx. A heterogeneous electron transfer constant between TTF and the MPA-GOx bioelectrode of 1.25 s(-1) was calculated. The enzyme loading, the amount of TTF on the electrode, the applied potential and the pH were also optimized. A good repeatability of the measurements with the TTF-GOx-MPA-AuE biosensor was demonstrated, with no need of pretreatment of the modified electrode. No significant changes in the slope value for the glucose calibration graph were found after 5 days when working with the same biosensor. An apparent Michaelis-Menten constant of (13.9 +/- 0.5) mM, and a limit of detection for glucose of 3.5 x 10(-6) M were obtained. Moreover, the TTF-GOx-MPA-AuE also performed well in the flow-injection mode and in the analysis of glucose in real samples.