A method, called the ac-TPD method, was presented for calculating the density distribution function of the adsorption enthalpy of ammonia by utilizing the temperature-programmed desorption (TPD) spectrum of ammonia from solid catalysts measured under complete adsorption equilibrium conditions. TPD spectra were measured using a gas stream containing ammonia, at a relatively large concentration compared to the amount of ammonia desorbed from the catalyst. Under these conditions, the concentrations of ammonia both inside the catalyst particles and in the gas phase are kept almost constant during the experiment. This situation meant that the amount of ammonia remaining on the catalyst was at an equilibrium state and dependent on the temperature and concentration of ammonia. Therefore, it could be assumed that the adsorption equilibrium of ammonia on acid sites can be expressed by a Langmuir-Hinshelwood type equation, which has an adsorption enthalpy and a pre-exponential adsorption equilibrium constant corresponding to the strength of an acid site. The overall TPD spectrum is represented by the sum of the spectra from acid sites with different acid strengths. Under these assumptions, the desorption temperature was related to the adsorption enthalpy of ammonia, and a relationship was derived between the overall TPD spectrum and the density distribution function of the adsorption enthalpy. The distribution function obtained by this method can be used for evaluating acidic properties of solid catalysts. The heat flux required for the desorption of ammonia from acid sites during the TPD experiment was calculated using the obtained distribution. The value was found to agree well with that directly measured by a differential scanning calorimetric apparatus. The density distribution of the adsorption enthalpy (Q) was compared to the distribution of activation energy for desorption of ammonia (E) by a method, called dc-TPD, which was reported previously by us. This comparison gave the relationship between Q and E, and it was found that the Q value was smaller than the E value by about 15 kJ mol(-1).