This review was aimed to evaluate health economic models used in evaluations of different treatment strategies in spondyloarthritis (SpA). Model-based health economic evaluation studies are increasing and complex models with short-term and long-term horizon are applied to investigate the cost-effectiveness of SpA treatments. The objective of this study was to carry out a systematic review of the evolution of health economic models used in the treatment of SpA. Electronic searches within MEDLINE and EMBASE were carried out using a predefined search strategy. Inclusion and exclusion criteria were used to select relevant studies. Data on country, intervention, evaluation perspective, type of model, time horizon, types of costs and effectiveness measurement were extracted. Eighteen models were described in 22 publications, of which 81.8% were European. Study perspectives included the societal (n=6), healthcare system and payer (n=14), or patient and government (n=1). Time horizon ranged from 52 weeks to lifetime. Markov model was the most frequently used model, only one individual patient simulation models accounting for uncertainty in multiple parameters was reported. Most studies compared different biologics (including different TNFi/biosimilar and IL-17A antibody) with conventional care (NSAIDs) because of the high prize. Only half of studies took indirect costs into account. Modeling is of importance in health economic evaluations of SpA treatment. Long-term costs especially indirect costs should be considered when comparing different treatment alternatives in order to provide more information for policy makers and clinicians.