The ponds and lakes in the industrial area represent intricate ecosystems that play a direct role in shaping the habitats of human populations. Monitoring the water quality in urban areas is very crucial. Turbidity (TUB) is a vital parameter in water quality, intricately linked to the penetration of underwater light, and it exerts an impact on primary productivity. The purpose of this work is to propose a neural network (NN) model with Adam optimizer to predict the TUB of lake using remote sensing. The correlation is obtained in between the reflectance values of Landsat 8 and measured TUB values using the neural network. The proposed NN consists of input, two hidden, and output layers. The reflectance values of six Landsat 8 bands are given as input to the model. The NN is trained with 191 parameters and the weights are adjusted by calculating moving average of gradient and moving average of squared gradient to reduce the error. The proposed model is compared with linear regression, random forest and k-NN regression model. The results indicate that the proposed neural network model has demonstrated better performance when compared to the alternative models.