ABSTRACT We present the results from a systematic search and characterization of ionized outflows in nearby galaxies using the data from the second Data Release of the Mapping Nearby Galaxies at Arecibo Point Observatory (MaNGA) Survey (DR2; >2700 galaxies, z ≤ 0.015). Using the spatially resolved spectral information provided by the MANGA data, we have identified ∼5200 H α-emitting regions across the galaxies and searched for signatures of ionized outflows. We find evidence for ionized outflows in 105 regions from 103 galaxies, roughly 7 per cent of all the H α-emitting galaxies identified in this work. Most of the outflows are nuclear, with only two cases detected in off-nuclear regions. Our analysis allows us to study ionized outflows in individual regions with star formation rates (SFRs) down to ∼0.01 M⊙ yr−1, extending the ranges probed by previous works. The kinematics of the outflowing gas is strongly linked to the type of ionization mechanism: regions characterized by low-ionization emission region emission (LIER) host the outflows with more extreme kinematics (FWHMbroad ∼ 900 km s−1), followed by those originated in active galactic nuclei (550 km s−1), ‘Intermediate’ (450 km s−1), and star-forming (350 km s−1) regions. Moreover, in most of the outflows we find evidence for gas ionized by shocks. We find a trend for higher outflow kinematics towards larger stellar masses of the host galaxies but no significant variation as a function of star formation properties within the SFR regime we probe (∼0.01–10 M⊙ yr−1). Our results also show that the fraction of outflowing gas that can escape from galaxies decreases towards higher dynamical masses, contributing to the preservation of the mass–metallicity relation by regulating the amount of metals in galaxies. Finally, assuming that the extensions of the outflows are significantly larger than the individual star-forming regions, as found in previous works, our results also support the presence of star formation within ionized outflows, as recently reported by Maiolino et al. (2017) and Gallagher et al. (2018).