Accurate path loss predictions are paramount for the optimization of future 6-G and 7-G in-vivo Wireless Nanosensor networks iWNSN. In this research, we focus on the critical aspect of in-vivo channel path loss prediction using the machine-learning approach for cardiac health monitoring. We initially present the theoretical model used for predicting and calculating the total path loss due to absorption and spreading loss. Using these models we present the numerical simulation results using the data collected from the online repository. We then use machine learning models such as Artificial Neural Network, Polynomial Regression, and Gradient Boosting algorithms to predict the total path loss in in-vivo communication channels and compare the performance of these models with the analytical models. The inference is that the gradient boosting model is outperforming compared to ANN and polynomial regression models with R2 scores of 0.9996, 0.9817, and 0.9547 respectively. We conclude from this work that machine learning models could be used for predicting path loss prediction in iWNSNs.