The service request scheduling problem in Mobile Edge Computing (MEC) often overlooks user mobility, request heterogeneity, and server load distributions, resulting in increased latency and energy consumption. To address these challenges, we introduce the Mobility-Aware Edge Service Scheduling with Request Heterogeneity and Server Load Balancing (MESS-HL) problem, considering user mobility, data volume changes of service requests, and load balancing. By proving and solving the \(\mathcal{N}\mathcal{P}\) -hard MESS-HL problem, we propose a novel heuristic simulated annealing algorithm, MHLSA. It integrates user mobility trajectories, heterogeneous requests, and server load status as heuristic information into both initialization and perturbation to find an approximately optimum solution. Comprehensive experiments on two real-world datasets show that MHLSA achieves an average improvement of 31.2% in response time, 62.2% in energy consumption, and 48.9% in load balancing over existing methods.