Nearly 5.3% of the national income of Bangladesh comes from fish. Fishes are the significant natural essentials that help to grow national income, nutrition, reduce the unemployment problem of a country and also earn foreign currency. Furthermore, it's a great source of low cost, high protein and other health beneficiary nutrients comparative to red meat. Nonetheless, to fulfill the expected demand for fish, the existing system and conventional fish farming has been failed to raise the amount of fish needed for the growing population. This paper analyzed the water quality parameters standards for the suitability of fish farming and the causes of fish diseases affected by the parameters through collected ponds data from the different areas of Bangladesh. Several machine learning algorithms have been compared for accuracy for the significance water level and error rate. Logistic regression has been fitted better to train and test part. The prediction has been done to find out whether the new pond's water quality is suitable for fish farming with respect to the value of quality parameters. An empirical IoT based system design has been given to comparing the prediction in the future. Moreover, this research also analyzed the feasible environment parameter and standards for fish growth, the reason, and risk for fish death as well as the growth rate of fish by monitoring the quality parameters of water for fish.