owing to the high complexity and the non-linearity of ecosystems and the rapid development of technology in spatial data survey, soft computation is more and more widely used in ecological models. Soft computation is a broad field which includes cellular automata (CA), individual based models and box-based model with respects to paradigms, and artificial neural networks (ANN), fuzzy logic (FL), genetic algorithm (GA), chaos theory and rules-based method with respects to techniques. This paper concentrates on the uses of CA paradigm and rule-based technique in ecological models. Application cases include population dynamics modeling, algal blooms forecasting and habitat simulation.