Agriculture is a vital economic sector in many regions and depends heavily on three-phase induction motors for irrigation pumps, threshers and other farm machinery, making their reliable and uninterrupted operation critical to crop yield and farmer livelihoods. Three-phase induction motors are used widely in industrial and agricultural applications. But these motors are prone to electrical faults like overcurrent, undervoltage and single-phasing. This causes insulation deterioration and unplanned shutdown. Conventional protection depends on fixed threshold relays with no remote supervision or prediction capabilities while cloud dependent Internet of Things systems are impractical in the rural environments without internet infrastructure. This paper proposes a LoRa and TinyML integrated framework which provides real time monitoring, protection, remote control and predictive maintenance of three-phase induction motors in constrained infrastructure environments. The system employs an ESP32 microcontroller, PZEM-004T voltage and current monitoring module, current transformer coil sensors, Reyax RYLR896 LoRa transceivers at 867 MHz and a relay-contactor protection mechanism. The entire process from detecting faults to isolating the motor and TinyML inference thus executes on local edge nodes. The ESP32 on the motor end acts as the Wi-Fi Access Point that supports an operator dashboard on a Local Area Network. The proposed system was validated through MATLAB/Simulink simulation and experimental implementation on a three phase, 2 HP, 415 V, 3.3 A, 50 Hz induction motor.