Snake robots under environments with avoidance constraints hardly track routes such that traditional guidance approaches become infeasible. Motivated by it, this work presents a line-of-sight-guidance virtual-snake-robot integrated scheme for snake robots to track paths and obey collision avoidance regulations. Main contributions include the avoidance regulation-compliant kinematics guidance law and finite-time-converged adaptive dynamic controller. In kinematics, a virtual leader is used to optimize nondifferentiable inflection into smooth to provide a jitter-free direction. Considering avoidance constraints, corresponding guidance strategies are innovatively designed for head-on, overtaking, and crossing scenarios to ensure the safety and autonomy of path tracking. In terms of dynamics, maneuvering control is executed by a finite-time disturbance observer-based event-triggered terminal sliding-mode controller. The unmeasured errors and uncertainties are accurately compensated, and input event-triggering rules can reduce the actuator burden. Finally, the stability analysis proves system errors' finite-time convergence, and result validations confirm the performance superiority and feasibility of this method.