Abstract A major international inter-comparison exercise on strain-controlled thermo-mechanical fatigue (TMF) has been undertaken to validate a new European Code of Practice for TMF Testing and to provide underpinning information for an ISO Standard. This paper focuses on (a) distribution of samples of Nimonic 90, (b) the establishment of a protocol for testing and reporting results, and (c) the analysis of the results. Participants in the inter-comparison exercise comprised eight inner-circle partners who primarily used test pieces which were all manufactured at the same workshop, albeit of three different test piece geometries, and ten outer-circle participants who manufactured their own test pieces, of their own in-house geometry. Each participant undertook three repeat in-phase (IP) tests and three repeat out-of-phase (OP) tests. The tests were conducted at temperatures cycling between 400 °C and 850 °C, with a strain range selected to give a failure life of approximately 1000 cycles, resulting in a stress range of up to ∼1000 MPa. The testing conditions were chosen following a preliminary evaluation of critical testing parameters. Results from solid circular and solid flat test piece geometries, together with hollow tubular test pieces have been compared. The influence of temperature measurement using different types of temperature sensors has also been investigated. In-house repeatability has been assessed, together with inter-laboratory reproducibility. The results have been correlated with modulus and thermal expansion data for individual tests. Initially, the largest contribution to scatter in the results was attributed to human errors in reporting the results, compounded by computer assisted ‘cut-and-paste’ errors. Once these obvious discrepancies had been corrected, it was possible to use the data sets to point to some recommendations regarding testing procedures that can be incorporated into the Code of Testing Practice [Hahner P et al. Code of practice for thermo-mechanical fatigue testing – TMF-Standard-Work Package 6 Report, September 2005] and provide technical underpinning for the ISO Standard. The results and the procedures used for analysis are presented.