Currently DNA profiling methods only compare a suspect's DNA with DNA left at the crime scene. When there is no suspect, it would be useful for the police to be able to predict what the person of interest looks like by analysing the DNA left behind in a crime scene. Determination of the age of the suspect is an important factor in creating an identikit. Human somatic cells gradually lose telomeric repeats with age. This study investigated if one could use a correlation between telomere length and age, to predict the age of an individual from their DNA. Telomere length, in buccal cells, of 167 individuals aged between 1 and 96 years old was measured using real-time quantitative PCR. Telomere length decreased with age (r=-0.185, P<0.05) and the age of an individual could be roughly determined by the following formula: (age=relative telomere length -1.5/-0.005). The regression (R(2)) value between telomere length and age was approximately 0.04, which is too low to be use for forensics. The causes for the presence of large variation in telomere lengths in the population were further investigated. The age prediction accuracies were low even after dividing samples into non-related Caucasians, males and females (5%, 9% and 1%, respectively). Mean telomere lengths of eight age groups representing each decade of life showed non-linear decrease in telomere length with age. There were variations in telomere lengths even among similarly aged individuals aged 26 years old (n=10) and age 54 years old (n=9). Therefore, telomere length measurement by real-time quantitative PCR cannot be used to predict age of a person, due to the presence of large inter-individual variations in telomere lengths.