Abstract: River ice growth and melting conditions are of highly relevance for ice flood disaster warning system.Understanding the physical mechanism of river ice growth and melting is vital to determine the potential existence of chaotic characteristics within river ice.To fully understand the characteristics of intrinsic energy distribution in river ice,we selected the Mohe Station in the upper reaches of the Heilongjiang upstream as research area,and performed in-situ observations of river ice temperature and voltage and air temperature from 2014 to 2018.Chaos theory was used to analyze the chaotic characteristics and their evolution of this river ice system,and the linear and nonlinear transition times of the river ice system were determined based on combinations of chaos theory-based results with long-term observations.Our results show layer temperature/voltage exhibits a well-organized during growth period,but shows an obvious disorder during the melting period.Most of the maximum Lyapunov index are less than 0 during the freezing period,and a few of the maximum Lyapunov index larger than 0 mainly concentrated on between 14 and 16 at pm,which changed from negative to positive during the melting period of river ice,showing chaotic characteristics.The maximum Lyapunov index turning positive date is found to be closely associated with the average above 0℃ date,indicating a transition from linear to nonlinear heat conduction in river ice when the average temperature exceeds 0℃.The study reveals the temporal evolution of layer temperature during river ice formation and melting,thereby offering support for water resources management in cold regions.