This research addresses the significant challenge of unreliable wireless communication, which hinders the performance of ZigBee-based wireless sensor networks (WSNs) in precision agriculture. A dual-slope log-distance path loss model was developed to accurately predict signal propagation complexities in dense vegetative environments for improved wireless communication. The study was conducted on a cassava farm in Ondo State, Nigeria, characterized by vegetation heights of 1.8 meters, making it an ideal site for investigation. A systematic methodology was employed, incorporating radio frequency measurements in both line-of-sight and non-line-of-sight conditions. This involved deploying two XBee S2C modules operating at 2.4 GHz, with one designated as a coordinator and the other as a router. The collection of Received Signal Strength Indicator (RSSI) and throughput data occurred at 5 meter intervals, with variations in the router's orientation. Results revealed a maximum communication range of 70 meters under non-line-of-sight conditions, compared to 140 meters in line-of-sight scenarios, where the path loss exponent was determined to be 1.78. The path loss exponents for the cassava fields were found to be 2.55 and 4.25. The developed dual-slope path loss model showed a strong fit to additional empirical data from a separate cassava farm location, achieving a Mean Absolute Percentage Error (MAPE) of 3.30 % and an R-squared value of 0.94. Hence, this model offers a comprehensive framework for characterizing radio wave propagation in agricultural environments, enhancing data transmission reliability and energy efficiency in smart farming applications.