Optical satellite-derived Normalized Difference Vegetation Index (NDVI) is by far the most commonly used vegetation index value for crop monitoring. However, it is quite sensitive to the cloud, and cloud shadows and significantly decreases its usability, especially in agricultural applications. Therefore, an accurate and reliable cloud correction method is mandatory for its effective application. To address this issue, we have developed an approach to correct the NDVI values of each and every pixel of the image captured by the Sentinel-2A satellite of the European Space Agency's Copernicus Program. The Chhattisgarh region of India was selected to analyze the variation in NDVI value. The NDVI value of each pixel shows a slight decrease in the value because of clouds. Cloud probability was calculated using the S2cloudless package provided by Sentinel Hub. The cloud probability of a pixel implies how densely the cloud is present over that pixel, thus inversely affecting the NDVI values. To understand the relationship between cloud probability and NDVI, we did a pixel-to-pixel (>25 million pixels) comparison between clouded and non-clouded NDVI Images on two consecutive dates in June 2021. Our analysis shows that an increase in 0.1 Units of cloud probability corresponds to a decrease in 0.2668 units of NDVI value. We further validated this finding on images captured during the months of August and September 2021. The validation was done on more than 35k square meters of actual farm fields and the results indicate that the proposed method shows more than 60% improvement in the cloud-affected NDVI images.