Aptamers are an alternative to antibodies in their role as biorecognition elements in analytical devices. Aptamers are in vitro selected oligonucleotides that can bind with high affinity and specificity to a wide range of target molecules, such as proteins or other organic and inorganic molecules. They have been used in numerous investigations, as therapeutic or diagnostic tools and they have been recently employed in analytical chemistry, as immobilised ligands or in homogeneous assays. In this work the development of an aptamer-based biosensor (aptasensor) for C-reactive protein (CRP), an important clinical bio-marker, will be presented. CRP was the first acute-phase protein to be discovered (1930) and is a sensitive systemic marker of inflammation and tissue damage. It is produced by hepatocytes in response to the cytokine interleukin-6 and its plasma half-life is of 19 h. CRP is considered a marker of vascular inflammation and a risk factor for cardiovascular disease. It has also a prognostic value for patients with acute coronary syndrome. The average concentration of CRP in blood is 0.8 mg/l but levels may increase up to 500 mg/l in case of inflammation. An RNA aptamer specific for CRP has been coupled to a surface plasmon resonance transduction, for the development of a new generation of biosensors. The CRP-specific RNA aptamer is composed of 44 bases and the binding constant Kd was estimated to be 125 nM. Several immobilisation protocols have been studied and optimised. In particular, different spacers (polyT spacer composed of 20 thymines and a triethylene glycol spacer) have been attached to the 5’ end of the aptamer to study the effect of the vicinity of the oligonucleotide to the solid surface. Moreover, binding conditions such as binding buffer, ionic strength and pH, have been optimised. With the best working conditions, a detection limit for CRP of 0.005 mg/l was reached with good selectivity towards human serum albumin. This detection limit allows a high dilution (around 1000 times) of serum samples, in order to overcome matrix non-specific effects when analysing real samples.