Abstract Background Little is known about the response of the equine respiratory muscles to training. Objectives To measure an index of inspiratory muscle strength (IMSi) before and after a period of conventional exercise training ( phase 1 ) and inspiratory muscle training (IMT), comparing high‐load (treatment) and low‐load (control) groups ( phase 2 ). Study design Prospective randomised controlled trial. Methods Phase 1 : Twenty National Hunt Thoroughbred racehorses performed an inspiratory muscle strength test (IMST) twice on two occasions; when unfit at timepoint A (July), and when race fit at timepoint B (October). Phase 2 : Thirty‐five Thoroughbred racehorses at race fitness were randomly assigned into a high‐load (treatment, n = 20) or low‐load (control, n = 15) IMT group. The high‐load group followed an IMT protocol that gradually increased the inspiratory pressure applied every 4 days. The low‐load group underwent sham IMT with a low training load. The IMT was performed 5 days/week for 10 weeks. The IMST was performed twice on two occasions, timepoint B (October) and timepoint C (January). Conventional exercise training and racing continued during the study period. The peak IMSi values obtained from the different groups at timepoints A, B and C were compared using a Wilcoxon Signed Rank Test. Results Phase 1 : There was a significant increase in IMSi from timepoint A: 22.5 cmH 2 O (21–25) to timepoint B: 26 cmH 2 O (24–30) ( p = 0.015). Phase 2 : From timepoint B to C there was a significant increase in IMSi for the high‐load group 34 cmH 2 O (28–36) ( p = 0.001) but not the low‐load group 26 cmH 2 O (24–30) ( p = 0.929). The peak IMSi at timepoint C was significantly higher for the high‐load than low‐load group ( p = 0.019). Main limitations Single centre study with only National Hunt horses undergoing race‐training included. Conclusions In horses undergoing race training there is a significant increase in IMSi in response to conventional exercise training and high‐load IMT.