Sharing global channel information at base stations (BSs) is commonly assumed for downlink multi-cell precoding. In the context of massive multi-input multi-output (MIMO) systems where each BS is equipped with a large number of antennas, sharing instant fading channel coefficients consumes a large amount of resource. To consider practically implementable methods, we study in this paper interference reduction based on precoding using the large-scale fading coefficients that depend on the path-loss model and are independent of a specific antenna. We focus on the downlink multi-cell precoding designs when each BS is equipped with a practically finite number of antennas. In this operation regime, pilot contamination is not the dominant source of interference, and mitigation of all types of interference is required. This paper uses an optimization approach to design precoding methods for equal qualities of service (QoS) to all users in the network, i.e.,maximizing the minimum signal-to-interference-plus-noise ratios (SINRs) among all users. The formulated optimization is proved to be quasi-convex, and can be solved optimally. We also propose low-complexity suboptimal algorithms through uplink and downlink duality. Simulation results show that the proposed precoding methods improve 5% outage rate for more than 1000 times, compared to other known interference mitigation techniques.