Subsequent synthesis and detection using droplets as microreactors have shown promise in the development of novel materials and drugs because microreactors enable small-scale synthesis and detection of covalent/non-covalent intermolecular interactions. Self-organization exhibited by autonomous droplets under non-equilibrium conditions is beneficial for manipulating the sequentiality and selectivity of droplet coalescence because expensive equipment or elaborate techniques are not required with the autonomy of droplets. However, to our knowledge, selective coalescence caused by the collective motion of self-propelled droplets has not been demonstrated in inanimate systems. Here, we report sequentially selective coalescence based on the dynamic collective pattern of self-propelled droplets composed of ethyl salicylate (ES) or butyl salicylate (BS). When ES and BS droplets were placed on an aqueous sodium dodecyl sulfate (SDS) solution, the collective motion of droplets resulted in three stages of selective coalescence on the time development. Initially, coalescence was observed only between different types of self-propelled droplets. Subsequently, the formed droplets selectively coalesced with ES droplets. Finally, mature droplets merged with BS droplets. The sequentially selective coalescence was discussed from the dynamic pattern formation of swarming droplets and the collapse of the SDS monolayer at the o/w interface caused by the difference in Laplace pressure and the interfacial instability at the contact point between droplets. Thus, this study formulates a strategy of sequentially selective coalescence of droplets via the collective motion of non-identical self-propelled droplets, promoting a new type of powerful and efficient automation technology based on an autonomous inanimate manner of spatiotemporal pattern formation under non-equilibrium conditions for the droplet manipulation.