Researchers are focusing on methods to prevent diseases by pre-symptomatic diagnosis. The human cell reacts to pathogen attacks by releasing biosignatures, such as proteins. Detecting the concentration of these proteins in a biological sample may help monitor the state of health in a person. Traditionally, RNA and antibodies are used for protein detection. But they are difficult to obtain for the thousands of possible protein targets. The need for a new detection method gave rise to the idea of synthetic antibodies. Researchers are looking at designing antibodies that have high binding affinity for a target so as to detect any concentration of a target. There are several existing techniques for predicting the binding affinity between proteins and ligands, but the need to screen thousands of potential ligands requires a faster and efficient method. In this paper, a model is explored to predict the affinity using protein microarrays. Intensity data is obtained from the microarrays for different concentrations of a target protein. Assuming that the reaction is under equilibrium and that the concentration of the proteins stay constant irrespective of binding with ligands, a model is proposed to estimate affinity from the intensities.