Multiple regressions are a statistical term used in machine learning that learns more complex relationships between two or more variables. Time series analysis is a specialized learning method that analyzes the sequence of data collected for a particular target variable over a time period to derive the same target variable. The hydroponic farming system helps plants to grow by using water-based nutrients rather than soil. This paper discusses implementing multiple regression and time series analysis frameworks for hydroponic planting systems using rainwater. The rainwater collected carries bacteria and viruses and hence, the physicochemical properties of rainwater vary. Such water needs to be purified, tested, and then used for a hydroponic system. The multiple regression analysis is implemented to investigate the physicochemical composition of rainwater. The initial analysis shows a positive regression coefficient between As-P, Ni-Cu, Cr-Ca, and Pb-P elements. In contrast, the rest of the elements show a negative relationship and the R-Square value derives to be 0.45 which is relatively small. It suggests adding a few supplementary nutrients to improve the quality of water. Further, the regression coefficient for improved water is determined for each combination of elements. The combination Hg-active carbon-Selenium and Pb-P-Zn, Pb-P-Vitamin C, As-Ca-Se, As-P-Si, and As-Ca-Zn show negative correlation thereby calculating the average coefficient as 0.665 and R-Square appears to be 0.64 as comparatively better and considerable. Furthermore, the growth of the plant is observed and the time series analysis ARIMA is implemented to find ACF and PCAF for different lags to measure the performance of the plant. The forecasted values show the increase in plant growth and guide future decisions. Finally, the model is evaluated and the analysis shows that the plant’s growth is 60% better if the rainwater is properly purified and supplementary nutrients are added to the water.