In the field of ophthalmology in particular, numerous studies have been undertaken to develop and implement methods for the diagnosis and evaluation of various retinal illnesses such as Age-related Macular Degeneration (AMD) and Diabetic Retinopathy (DR). This paper aims at registration of Ultra-Widefield Fluorescein Angiography (UWF FA) retinal images. In the case of neurovascular illnesses like DR, UWF FA is the ultimate benchmark imaging modality because of the accuracy with which it portrays the retina’s neurovascular anatomy. Estimating a true set of matching corresponding pixels has always been a challenge for image registration. In this article, we formulate the alignment problem as an estimation problem using Expectation Maximization (EM) algorithm. With a given input of set of correspondence points and their distribution function, a posterior probability of correspondence can be estimated and maximized until convergence. A simple thresholding can then distinguish between the inliers and outliers which can recover the coherent pixel correspondence between the UWF FA retinal image pairs. The efficacy of the proposed algorithm has been quantitatively evaluated using the Information Ratio (IR) and the Mutual Information Ratio (MIR). While the IR describes characteristics unique to a single image, the MIR describes those that appear in several images. In terms of both IR and MIR, the proposed method surpasses the experimental results.