To predict natural events such as intense precipitation, flooding, and landslides, analyzing trends in rainfall patterns and conducting frequency analysis using the most suitable probability distribution function (PDF) is essential. The study aims to perform trend analysis and conduct frequency analysis after determining the best-fit PDF based on 35 years of annual rainfall data from 1988 to 2023 for nine stations in the Chittagong division, Bangladesh. Sen’s Slope Estimator and the Mann-Kendall trend model were used for trend analysis. Several PDFs, such as Gumbel, Generalized Extreme Value, Pearson Type-III, Log Pearson Type-III, Log-Normal, and Normal, were used, and return periods were estimated. Three goodness-of-fit tests—the Anderson-Darling, Chi-Square, and Kolmogorov-Smirnov—are performed to determine the best-fit PDF. All stations showed a declining trend, except CL325 (which exhibited an upward trend). Three of the stations showed statistically significant trend lines, while the remaining stations were not. Among the nine stations, the Generalized Extreme Value (GEV) distribution is the best fit for five, the Log-Normal for two, the Normal for one, and the Log Pearson Type-III for another, indicating that GEV is the most suitable for rainfall frequency analysis. For the return period, CL325 showed more variation in rainfall, whereas CL302 showed less variation. The findings of this study will enhance the understanding of current rainfall patterns and provide valuable insights into predicting future hydrological extremes, such as intense precipitation and flooding, in the region.