Adaptive sampling and planning in robotic environmental monitoring are challenging when the target environmental process varies over space and time.  I will first discuss a Monte Carlo tree search method which enables the robot to not only well balance the environment exploration and exploitation in space, but also catch up to the environmental dynamics that are related to time. This is achieved by incorporating multi-objective optimization and a look-ahead model-predictive rewarding mechanism. The method produces optimized decision solutions for the robot based on its knowledge (estimation) of the environment model, leading to better adaptation to environmental dynamics. Then I will discuss robot decision-making in uncertain and unstructured environments, such as in the scenario when strong winds and water flows cause robot stochastic behaviors. We explore the time-varying stochasticity of robot motion and investigate robot states' reachability, based on which we develop an efficient iterative method that offers a good trade-off between solution optimality and time complexity.