Using variables such as client size, complexity, risk, internal auditor, auditor size, and industry, empirical research explains price behaviour of audit fees, and often attributes this pattern to audit fee model. Our study helps to explain why the audit fee model does not fully explain the level of audit fees, because it does not take account of fee stickiness and the influence of previous fees. This paper examines the audit fee stickiness in Iranian firms, using panel data analysis. The sample is comprised of 79 publicly traded firms listed in Tehran Stock Exchange (TSE) using object oriented sampling method. Using data for 1386-1392, we find evidence that audit fees are sticky, i.e. they do not fully adjust to the levels predicted by a standard audit fee model. Audit fees also go up more quickly than they come down. We also find that the difference between upward and downward audit fee adjustments reduces in the second year and when longer periods of time are aggregated. Furthermore the results show that changing auditors has no effect on audit fee stickiness.