This research work studied the dissociation constants (pKa) of eight amines [N-(2- Aminoethyl)-1,3-propanediamine, Bis[2-(N,N-dimethylamino)ethyl] ether, 2- Methylpentamethylene diamine, N,N-Dimethyldipropylenetriamine, 3,3’-Diamino-Nmethyldipropylamine, 2-[2-(Dimethylamino)ethoxy]ethanol, 2-(Dibutylamino)ethanol, and N-Propylethanolamine] within a temperature range of 298.15K – 313.15K, using the potentiometric titration method. The thermodynamic quantities including the standard state enthalpy change (ΔH0) and the standard state entropy change (Δ𝑆0) for the dissociation process were determined via Van’t Hoff equation. The pKa values reflected the basicity of amines and results showed that all studied amines had a stronger basicity than methyldiethanolamine (MDEA) The pKa values of series of amines (25 compounds) relevant to CO2 capture were predicted based on the feedforward artificial neuron network (ANN) with the backpropagation algorithm. Eight parameters were used as the input data, and these parameters were divided into two categories: (a) molecular weight, critical pressure and critical pressure as inputs that were used to identify the compound; (b) temperature and physical properties as inputs including density, viscosity, surface tension and refractive index that were used to correlate pKa values. An optimized architecture of 8-5-7-1 was selected and predicted outputs were in a good agreement with targets, whose regression coefficient was 0.99424 and mean squared error for training, validation and test process was 2.20E-05, 0.0094 and 0.0078, respectively. To compromise the flexibility of the ANN model, the other architecture of 6-5-7- 1 which reduced density and viscosity as inputs was selected, and it had a regression coefficient was 0.99216 and mean squared error for training, validation and test process was 4.40E-05, 0.0045 and 0.0203, respectively.