The design logic and operating mechanism of recommendation algorithms are in line with the narrative concepts and laws of ideological and political education, which is the possibility for recommendation algorithms to bring new ideas to the narrative of ideological and political education. Introducing recommendation algorithms can optimize the human centered narrative logic of ideological and political education, promote two-way interactive narrative methods, and develop multidimensional narrative structures. However, as a new variable and risk catalyst in the narrative of ideological and political education, recommendation algorithms are gradually alienated under the influence of traffic and capital, which also brings new risks to the narrators, narrative objects, narrative content, and narrative fields of ideological and political education. Specifically, recommendation algorithms weaken the authority of narrators, affect the cognition of narrative objects, intensify the dissemination of low-quality content, and weaken the influence of narrative fields through “algorithm gatekeeping”“algorithm cocoons”“algorithm flooding” and “algorithm circles”. Therefore, in the process of narrative in ideological and political education, it is necessary to actively control recommendation algorithms, reshape the authority of narrators through criticism and construction, break through the “cocoon” of narrative objects through reflection and improvement, pay attention to regulation and guidance to purify the narrative content ecology, and strengthen the cultural construction of the narrative field through excavation and breaking through walls, so that recommendation algorithms can better empower the narrative of ideological and political education.