We sought a physiological marker of overtraining in horses, using commonly practised field and laboratory tests to allow early prediction and treatment of the syndrome. Thirteen Standardbred horses were trained as follows: phase 1 (endurance, 7 weeks), phase 2 (high intensity, 9 weeks) and phase 3 (overload, 18 weeks). In phase 3 the horses were divided into 2 groups: overload training (OLT) and control (C). The OLT group exercised at greater intensities, frequencies and durations than the C group. Overtraining occurred after 31 weeks and was defined as a significant decrease in treadmill run time to fatigue (RT) in response to a standardised exercise test (SET). Variables measured included: feed intake, bodyweight (BWT), resting haematology and plasma biochemistry and treadmill SETs to measure RT. The OLT group had a decrease in BWT after week 28 (P<.05) without a reduction in feed intake and a reduction in RT during the SET after 31 weeks. Signs persisted after 2 weeks of a reduced training load confirming overtraining. Haematology and biochemistry failed to detect any markers of overtraining. Although no physiological markers of overtraining were identified, empirical observations revealed that the behaviour of horses in the OLT group was different from those in the C group during the period of overtraining. This study reflects that a model of overtraining has been developed based on measurement of a reduction in performance; however, there were no consistent changes in haematology or serum biochemical values in association with the decrement in performance capacity.