This research investigates reliability improvement for demand-responsive feeder transit (DRFT) with simultaneous pick-up and drop-off using dynamic pricing scheme. The fare model adjusts charges based on arrival punctuality-compensating passengers when vehicles miss their time windows, with compensation proportional to delay duration. The flexible departure system is designed to enhance connection reliability. The optimization model maximizes operator profit while minimizing passenger costs, constrained by vehicle operating time, passenger time windows, capacity limits, and drop-off schedules. A multi-chain genetic algorithm solves this multi-objective routing problem. Case studies demonstrate: 1) Flexible scheduling outperforms fixed scheduling in connection reliability and operational profitability; 2) Dynamic pricing scheme surpasses fixed fares in connection reliability, profitability, and resource utilization. The integrated approach significantly enhances overall DRFT system reliability.