Abstract: Hydration and setting time of paste-like backfill slurry in the Gansu Province’s Jinchuan copper and nickel mine is slow, and the degree of segregation of coarse aggregate is high, seriously affecting the quality of cemented paste backfill. In this paper, by taking the unclassified tailings, waste rock and rod milling sand in Jinchuan’s No. 2 mining area as the experimental materials, and adopting the comprehensive test design method, the effects of different mass fraction, coarse aggregates and tailings-coarse aggregate ratio (mass ratio of unclassified tailings to coarse aggregate) on the setting performance, unconfined compressive strength and rheological properties of cemented paste backfill were studied. The experimental results show that the coarse aggregate's specific surface area and chemical composition (active MgO and CaO) in the unclassified tailings-coarse aggregate paste are the main factors influencing the setting time. Increasing the tailings-coarse aggregate ratio decreased the setting time of the paste backfill theory. Increasing the tailings-coarse aggregate ratio increased the yield stress of paste backfill slurry. With the increase in the tailings-coarse aggregate ratio, the plastic viscosity of paste backfill slurry (unclassified tailings-waste rock, unclassified tailings-waste rock-rod milling sand paste) increased. The unconfined compressive strength of the unclassified tailings-waste rock paste is better than that of the unclassified tailings-waste rock-rod milling sand paste. The shortest setting time and the best unconfined compressive strength (the unclassified tailings-waste rock paste, tailings-coarse aggregate ratio 5∶5) were reduced by 2.1 h, individually. They were also increased by more than 33% relative to the setting time, and unconfined compressive strength of the mine. Finally, the setting performance was optimized for single-objective and multi-objective regression. The multi-objective regression optimization showed that optimum setting time for the unclassified tailings-waste rock-rod milling sand paste was approximately 270 to 300 min, while for the unclassified tailings waste rock rod milling sand was approximately 10∶6∶6–10∶7∶7 and yield stress was about 167.0 to 169.0 Pa. The optimum setting time of the unclassified tailings-rod milling sand paste was found to be about 300–330 min for the single-objective regression, the unclassified tailings rod milling sand was approximately 10∶14–10∶16, and yield stress was about 164.0–167.0 Pa, which met the mine production requirements.