Recent observations reported hundreds of exoplanet systems composed of super-Earth mass planets in close-in orbits (e.g., Fabrycky et al. 2014; Weiss et al. 2018). While some of these planet systems are in chains of resonances (e.g., Mills et al. 2016; MacDonald et al. 2016; Gillon et al. 2017), most of them are not (Fabrycky et al. 2014; Winn & Fabrycky 2015). N-body simulations showed that planets are trapped in resonant chains through orbital migration induced by the planet-gas interaction (e.g., Terquem & Papaloizou 2007; Ogihara & Ida 2009). They cause orbital instability when the number of trapped planets is large, while their orbits are stable when the number is small (Matsumoto et al. 2012). Izidoro et al. (2017) found that both the fraction of the resonant systems and the number of planets in resonant chains are smaller in the observed systems than those by N-body simulations. We have investigated whether the orbital instability of planets in resonant chains is triggered by the long term mass evolution since the masses of planets and stars would evolve after disk gas depletion. We find that planets cause orbital instability even when the number of planets is smaller than the critical number estimated by Matsumoto et al. (2012). When the number of planets is slightly less than the critical number, the orbital instability can be triggered by a change of ≲10% in planetary mass. Such a planetary mass loss is likely to occur when a planetary envelope is lost via photoevaporation or core-powered mass loss. The observed properties of the resonant chain can be explained by the orbital instability triggered by the mass evolution of planets and stars.